内容

Contents

详细内容

Detailed contents

介绍

Introduction

“历史是普遍的、基本的，”一位历史学教授在一次教职工会议上说，“它关乎所有时间、所有地点的所有人和每一件事。” “是的，”他的语言学同事说道，“但是如果没有语言，你如何记录和解读历史呢？” 事实上，很难想象如果没有语言，没有通过复述故事和传奇将发生的事情记录从一代传给下一代的手段，甚至在书面记录出现之前，历史怎么会存​​在。 人类历史（和史前史）的大部分内容包括开发和改造各种工具以满足广泛的需求：想想轮子、动物驯养、蒸汽机、电脑和互联网。如果没有语言，这些工具以及所有其他工具的开发和完善就不可能完成。

“History is universal and basic,” a history professor said during a faculty meeting, “It’s about every event that involves all people at all times and in all places.” “Yes,” observed his colleague from linguistics, “but how would you record and interpret that history without language?” Indeed, it is hard to imagine how there could even be history without language, without a means to pass a record of what has happened from one generation to the next through retold stories and sagas, even before written records. Much of the history (and prehistory) of the human species consists of the development and adaptation of various tools to meet a broad range of needs: think of the wheel, the domestication of animals, the steam engine, computers, and the internet. The development and refinement of these and all other tools could not have been accomplished without language.

人类的自我意识和抽象思维能力即使不依赖于语言，也得益于语言。传递复杂信息、讨论事件意义和替代行动可能产生的结果、分享感受和想法的能力——如果没有语言，所有这些都是不可能的。语言的起源晦涩难懂，但考古记录表明，语言交流大约在 20 万年前就出现了。个人为自己塑造世界和使用语言交流的能力可能是人类物种最有利的进化适应能力。

The human capacity for self-awareness and abstract thought is facilitated by language, if not dependent upon it. The ability to transfer complex information, to discuss the meaning of events and possible outcomes of alternative actions, to share feelings and ideas – all these are impossible without language. The origins of language are shrouded in obscurity, but archaeological records suggest that communication with language emerged about 200,000 years ago. The ability of an individual to model the world for him/herself and to communicate using language was probably the single most advantageous evolutionary adaptation of the human species.

定义语言

Defining language

可以想象，语言的精确定义并不容易，因为语言现象复杂且面面俱到。Finegan 和 Besnier ( 1989 ) 认为，我们可以将语言定义为由元素和原则组成的有限系统，它使说话者能够构造句子来完成特定的交际工作。系统的一部分使说话者能够生成和解释符合语法的句子，这部分称为语法能力。它包括哪些语音属于特定语言，以及这些语音可以或不可以如何组合在一起的知识。语法能力还包括了解一种语言中不同语音序列所表示的含义，以及如何将这些意义单位组合成单词、短语和句子。语法能力使英语使用者能够将 21 个语音组合在一起，听起来像“狗把猫赶上树”，并允许另一个英语使用者理解狗、猫、和树是什么，追逐是什么，哪个方向是向上。此外，语法能力使这些英语使用者能够理解是狗在追逐，是猫爬上了树。当然，这不仅适用于英语。语法能力同样有助于理解所有人类语言。

As one can imagine, a precise definition of language is not easy to provide, because the language phenomenon is complex and has many facets. Slightly modifying a definition provided by Finegan and Besnier (1989), we might define language as a finite system of elements and principles that make it possible for speakers to construct sentences to do particular communicative jobs. The part of the system that allows speakers to produce and interpret grammatical sentences is called grammatical competence. It includes the knowledge of which speech sounds are part of a given language and how they may and may not be strung together. Grammatical competence also includes knowing the meanings signified by different sound sequences in a language and how to combine those units of meaning into words, phrases, and sentences. Grammatical competence is what allows a speaker of English to string together twenty-one sounds that sound something like “The dog chased the cat up the tree” and allows another speaker of English to understand what dogs, cats, and trees are, what chasing is, and which way is up. Further, grammatical competence is what allows these speakers of English to share the understanding that it was the dog doing the chasing and that it was the cat that went up the tree. Of course this does not apply only to English. Grammatical competence contributes similarly to comprehension in all human languages.

但人们使用语言不仅仅是为了传达语法句子的字面意义。“狗把猫赶上树”这个句子可能用于完成各种各样的工作：讲述故事的一部分，向狗的主人抱怨，帮助猫的主人找到他的宠物。定义的第二部分“做特定的交流工作”是指交际能力。人们用语言做的最频繁的“工作”就是与他人交流。

But people use language to do far more than just communicate the literal meanings of grammatical sentences. The sentence “The dog chased the cat up the tree” might be used to accomplish a wide variety of jobs: to narrate part of a story, to complain to the dog’s owner, to help the cat’s owner find his pet. The second part of the definition, “to do particular communicative jobs,” refers to communicative competence. The most frequent “job” that people do with language is communicate with other people.

如果没有交际能力，语法能力对于人际交往几乎毫无用处。事实上，语言的很多实际使用根本不是在句子中，而是在比句子大或小的话语单位中，有些是语法的（在形式语言学中使用的技术意义上），有些不是。为了有效，说话者必须将语法能力与如何根据当前目的和语境适当使用语法句子（和其他语言结构）的知识结合起来。两者结合起来构成了交际能力。交际能力——语法能力中包含的知识加上使用该知识完成各种交际工作的能力——构成了语言。

Grammatical competence is almost useless for human interaction without communicative competence. In fact, a lot of the actual use of language is not in sentences at all, but in discourse units larger and smaller than sentences, some grammatical (in the technical sense used in formal linguistics), some not. To be effective, speakers have to combine grammatical competence with the knowledge of how to use grammatical sentences (and other pieces of linguistic structure) appropriately for the purpose and context at hand. The two taken together comprise communicative competence. Communicative competence – the knowledge included in grammatical competence plus the ability to use that knowledge to accomplish a wide range of communicative jobs – constitutes language.

语言的普遍属性

Universal properties of language

经过数千年的进化，人类的声道已经足够灵活，能够发出各种可辨别的声音，并且能够感知这些声音之间的差异。但最重要的是，人类已经发展出使用这些声音来传达意义的能力。没有人知道这是如何发生的。也许，为各种其他适应目的而进化的心理能力（如精细运动手眼协调）被“重新利用”来支持复杂的符号和交流系统。也许某些心理能力专门用于语言，并随着人类交流的日益复杂而逐渐进化。或者，一旦达到一定的神经和认知复杂程度，大脑的突触就会“重组”，从而使语言的发展成为可能。

Over thousands of years of evolution, the human species developed a vocal tract flexible enough to produce a wide range of distinguishable sounds and the ability to perceive differences among those sounds. But most important, the human species developed the ability to use these sounds in systems which could communicate meaning. No one knows just how this happened. Perhaps mental capacities that had evolved for a variety of other adaptive purposes (like fine motor hand–eye coordination) were “re-purposed” to support a complex symbolic and communicative system. Perhaps some mental capacities are exclusively dedicated to language and evolved more gradually along with the increasing complexity of human communication. Or perhaps once they reached a certain level of neurological and cognitive complexity, the synapses of the brain “reorganized” themselves, making the development of language possible.

尽管语言在很多方面都存在差异，但它们都是由相同的遗传信息形成的，它们都以基本相同的方式被大脑处理，而且毫不奇怪，它们都具有某些基本的“设计特征”和结构特征，使它们能够以它们的方式发挥作用。例如，尽管不同的语言使用不同的声音组合，但它们的声音是根据一些原则组织和组合的。如果语言没有共同的、普遍的特征，我们就会认为语言的声音及其组合会随机变化。相反，语言的声音及其组合是有限的和系统的。同样，所有语言在如何将单词组合成短语和句子方面都遵循类似的限制。

Although languages differ in many ways, they are all made possible by the same genetic information, they are all processed by the brain in basically the same ways, and, not surprisingly, they all share certain fundamental “design features” and structural characteristics that enable them to work the way they do. For example, although different languages use different sets of sounds, their sounds are organized and combined according to just a few principles. If there were no shared, universal features of language, we would expect the sounds of languages and their combinations to vary randomly. Instead, the sounds of languages and their combinations are limited and systematic. Likewise, all languages follow similar constraints on how they can combine words into phrases and sentences.

理解和解释所有语言的普遍属性以及不同语言之间的差异属性是语言学家的基本工作。

Understanding and explaining the properties which are universal to all languages – as well as those which vary across languages – is the fundamental job of the linguist.

模块化

Modularity

大多数语言学家认为语言是一个模块化系统。也就是说，人们使用一组子系统（或模块）以协调的方式产生和解释语言。每个模块负责整个工作的一部分；它将其他模块的输出作为输入，并将自己的输出分配给其他模块。神经语言学研究表明，大脑的不同区域与语言处理的不同方面相关，正如以下章节所示，将语言划分为模块极大地方便了语言分析。

Most linguists believe that language is a modular system. That is, people produce and interpret language using a set of component subsystems (or modules) in a coordinated way. Each module is responsible for a part of the total job; it takes the output of other modules as its input and distributes its own output to those other modules. Neurolinguistic studies show that different regions of the brain are associated with different aspects of language processing and, as the following chapters show, dividing language into modules facilitates linguistic analyses greatly.

有些模块长期以来一直是语言学的核心。语音学是关于语音的产生和解释。音系学研究语言整体以及个别语言中原始语音的组织。更大的语言单位是形态学领域，研究单词的结构——以及句法领域，研究句子的结构。与这些模块交互的是词汇，即语言元素及其意义和结构属性的储存库。近几十年来，哲学家们已经发展了语义学的正式研究（对字面意义的详细分析），语言学也已将语义学纳入并作为语言的另一个模块。最近，话语（超越句子的语言组织）已被大多数语言学家视为语言的另一个重要子系统。

Some modules have been central to linguistics for a long time. Phonetics is about production and interpretation of speech sounds. Phonology studies the organization of raw phonetics in language overall as well as in individual languages. Larger linguistic units are the domain of morphology, the study of structure within words – and of syntax, the study of the structure of sentences. Interacting with these modules is the lexicon, the repository of linguistic elements with their meanings and structural properties. In recent decades, philosophers have developed the formal study of semantics (the detailed analysis of literal meaning), and linguistics has incorporated and added semantics as another module of language. Still more recently, discourse – organization of language above and beyond the sentence – has been recognized by most linguists as another important subsystem of language.

离散性

Discreteness

语言的每个模块都处理一些离散的语言单位（音素、词素、单词、短语、句子、话语）的特征、分布和协调。离散性是语言的另一个属性，它将连续的声音或意义空间划分为离散的单位。人类可以发出的声音范围是连续的，就像滑哨一样。例如，你可以从高音“长e ”（如feed）一直滑到低音“短a ”（如bat）。但所有语言都将连续的声音空间划分为离散的类别，就像大多数西方音乐将连续的音高范围划分为音阶中的离散步骤一样。在一种语言中离散的声音在另一种语言中可能不是离散的。例如，在英语中，我们区分 [a]（“短 a ”）和 [ ɛ ]（“短e ”），因此pat和pet是不同的词。德语中情况并非如此，所以德语使用者很难听出pet和pat之间的区别。同时，德语有一个元音类似于英语的“长a ”，但唇形是圆的，拼写为ö，称为“ o -umlaut”。类似于英语“长a ”的元音与这个圆唇元音之间的区别导致了Sehne（“tendon”）和Söhne （“sons”）之间含义上的差异。对于德语使用者来说，这种区别就像英语使用者区分pet和pat一样容易，但对于英语使用者来说却很难。究竟 什么是离散的，在不同语言中是不同的，但所有语言都具有离散性。

Each module of language deals with the characterization, distribution, and coordination of some discrete linguistic unit (phonemes, morphemes, words, phrases, sentences, utterances). Discreteness, another property of languages, divides the continuous space of sound or meaning into discrete units. The range of sounds that human beings can make is continuous, like a slide whistle. For example, you can slide from a high “long e” sound (as in feed) all the way down to a low “short a” sound (as in bat) in one continuous glide. But all languages divide that continuous space of sound into discrete categories, just as most western music divides the continuous range of pitch into discrete steps in a scale. Sounds that are discrete in one language may not be discrete in another. In English, for example, we distinguish [a], “short a,” from [ɛ], “short e,” so that pat and pet are different words. The same is not true in German, so German speakers have trouble hearing any difference between pet and pat. At the same time, German has a vowel that is like the English “long a,” but with rounded lips, spelled ö and called “o-umlaut.” The distinction between the vowel that is like English “long a” and this rounded vowel is responsible for the meaning difference between Sehne (‘tendon’) and Söhne (‘sons’). This distinction is as easy for German speakers as the pet and pat distinction is for English speakers, but it is hard for English speakers. Precisely what is discrete varies from one language to another, but all languages have the property of discreteness.

离散性还体现在语言的其他模块中，例如含义。色谱就是一个明显的例子。颜色变化是一个连续体——红色从红橙色渐变为橙色，再到黄橙色，最后变成黄色，依此类推。但是，所有语言都将色谱划分为离散的类别，尽管语言在将连续体划分为词语方面有所不同。有些语言只有两个基本颜色词，大致意思是“亮”和“暗”；其他语言则添加了红色、黄色和绿色，而其他语言（包括英语）则发展出了更多颜色的词语。同样，虽然爱斯基摩人有数百种描述雪的词语的说法可能言过其实，但生活在遥远北方的美洲原住民的语言确实比那些为了满足生活在温暖气候中的人们的需求而发展的语言区分了更多种类的雪。同样，美式英语也有一系列用于描述不同类型汽车的词语（例如轿车、运动型多用途车、小型货车、敞篷车、旅行车、跑车），这些词语与汽车在该文化中的重要性有关。

Discreteness also shows itself in other modules of language, such as meaning. The color spectrum is a clear example. Color variation is a continuum – red shades through red-orange to orange to yellow-orange to yellow and so on through the spectrum. But all languages divide the color spectrum into discrete categories, although languages differ in how they divide that continuum into words. In some languages there are only two basic color terms, roughly meaning ‘light’ and ‘dark’; others add red, yellow, and green, whereas still others, including English, have developed words for many more colors. Likewise, although the claim that Eskimos have hundreds of terms for snow may be overstated, the languages of Native Americans living in the far north do distinguish more kinds of snow than do languages which have developed to meet the needs of peoples living in warmer climates. Similarly, American English has a range of words for different types of automotive vehicles (sedan, sports utility vehicle, minivan, convertible, wagon, sports car, for example) related to the importance of the automobile in that culture.

语言由独立的声音、单词、句子和其他发声单位组成。从听觉上讲，声音和单词是相互融合的。（如果你成年后尝试学习第二语言，你就会知道区分以正常对话速度说出的单词有多难。）令人惊讶的是，只有几周大的婴儿能够区分他们家乡语言中非常相似的声音，并在很小的时候就区分出他们正在学习的语言的声音和其他语言的声音。此外，生命的第一年或第二年的孩子无需任何指导就能学会从语音流中挑选出单词。我们听到的语音是一系列单独的声音、单词和句子，这实际上是一项令人难以置信的成就（更令人难以置信的是，我们是在瞬间和无意识中做到这一点的它）。

Language is composed of separate sounds, words, sentences, and other utterance units. Acoustically sounds and words blend into each other. (If you have tried to learn a second language as an adult, you know how hard it can be to separate words spoken at a normal conversational pace.) Remarkably, babies only a few weeks old are able to distinguish even closely related sounds in the language of their home from each other and to distinguish the sounds that belong to the language they are learning from the sounds in other languages at a very early age. Furthermore, children in the first year or two of life learn to pick out words from the stream of speech with no instruction. The fact that we hear speech as a sequence of individual sounds, words, and sentences is actually an incredible accomplishment (and all the more incredible for how instantaneously and unconsciously we do it).

选区

Constituency

所有语言都将这些基本的离散单位组织成成分，即语言单位组，这样更复杂的单位就可以进入结构中，而更简单的单位也可以进入结构中。所以我们可以用英语说，“她坐了下来”，“那个聪明的女人坐了下来”，“那个身材高挑、黑发、穿着鲜红色毛衣戴着珍珠项链的聪明女人坐了下来”。每个斜体短语都构成一个名词短语（在这些例子中是句子的主语）；名词短语可以像第一个句子中的代词一样简单，也可以通过使用形容词和介词短语修饰名词而变得更复杂。由成分组成使语言具有结构和灵活性的平衡。成分可以被其他成分替换，但不能用一系列不是成分的单词替换成分。所以你不能用穿着鲜红色毛衣的聪明人替换她（ “穿着鲜红色毛衣的聪明人坐下”不起作用）。成分可以移动，但只能移动完整的成分。她非常聪明是可能的，而且非常聪明，她也是，但不是聪明，她非常。

All languages organize these basic discrete units into constituents, groups of linguistic units which allow more complex units to enter structures where simpler ones are also possible. So we can say in English, “She sat down,” “The smart woman sat down,” “The tall, dark-haired, smart woman with the bright red sweater and pearl necklace sat down.” Each italicized phrase constitutes a noun phrase (which is the subject of the sentence in these examples); a noun phrase can be as simple as a pronoun as in the first sentence, or it can be made more complex by modifying the noun with adjectives and prepositional phrases. Being composed of constituents gives language a balance of structure and flexibility. Constituents can be replaced by other constituents, but you can’t replace a constituent with a series of words that is not a constituent. So you can’t replace she with smart with the bright red sweater (“Smart with the bright red sweater sat down” doesn’t work). Constituents can be moved, but you can only move a complete constituent. She is very smart is possible and so is Very smart, she is, but not Smart, she is very.

递归和生产力

Recursion and productivity

由成分组成使得语言具有递归性。递归是系统的一种属性，允许重复应用某个过程。在语言中，我们可以结合成分产生无限多种长度不定的句子。例如，英语中的协调允许我们将两个或多个相同类型的成分组合在一起。我们可以将短句（如“他很高” ）扩展为较长的句子（如“他又高又强壮又英俊又体贴又善于倾听”等） ……或者无限嵌入从句来修饰名词短语，如“这是一只老鼠，它啃了杰克建造的房子里的奶酪”。

Being composed of constituents allows languages to be recursive. Recursion is a property of systems which allows a process to be applied repeatedly. In language we can combine constituents to produce an infinite variety of sentences of indefinite length. For example, coordination in English allows us to combine two or more constituents of the same type together. We can expand a short sentence like He was tall into longer sentences like He was tall and strong and handsome and thoughtful and a good listener and … or infinitely embed clauses to modify noun phrases, as in This is the mouse that nibbled the cheese that lay in the house that Jack built.

语言的递归性具有深远的意义。这意味着没有人可以通过记住该语言的所有句子来学习语言；相反，他们必须学习创建和组合该语言成分的系统。人类的大脑是有限的，但语言的递归性意味着通过学习一门语言，我们能够产生和理解无限数量的句子。语言的这种非有限性源于其生产力。即使一个人试图记住曾经说过的所有句子，他总是可以通过语言的递归规则添加另一个修饰词——（一座设计精美、结构精巧、位置优越的新建筑……）或将一个句子嵌入另一个句子中，一遍又一遍（他说她说我说他们相信你告诉我们……）。由于语言对这些递归过程的使用没有限制，因此所有语言都具有无限生产力的潜力。

The recursiveness of language has profound implications. It means that no one can learn a language by memorizing all the sentences of that language; instead, they must learn the system for creating and combining constituents in that language. The human brain is finite, but the recursive property of language means that by learning a language we are capable of producing and understanding an infinite number of sentences. This nonfinite quality of language is due to its productivity. Even if one were to attempt to memorize all the sentences ever uttered, one could always add another modifier – (A great big huge beautifully designed, skillfully constructed, well-located new building …) or embed one sentence within another, over and over again (He said that she said that I said that they believe that you told us that …) through the recursive rules of the language. Since languages place no limits on the use of these recursive processes, all languages are potentially infinitely productive.

新词（新造词）也体现了语言的生产力，它们在整个历史和社会中都出现过。当人们第一次听到一个词时，他们常常会问：“这是一个词吗？”如果他们问语言学家，答案很可能是“现在是一个词了”。如果新词是根据其语言的形态和音位规则形成的，并且在上下文中可以理解，那么它就是一个真正的词，即使它在词典中找不到。以最近创造的单词bling为例，意思是“闪亮的珠宝”。它在音位上是合乎规范的（在英语中， bl可以放在音节开头，ng [ ŋ ] 可以放在音节结尾）。这个词已经在主流公众中流行起来，现在是一个真正的词。大多数这些自发创造的词语（受特定语境启发，通常被称为俚语）使用频率不高，因此从未被收录到词典中，但有些词语 确实成为词汇的一部分（并被收录到一些更新的词典中），因为它们满足了新的需求。创造新词是一种富有成效的过程，语言通过这种过程发生变化，以满足使用者不断变化的交流需求。

Productivity in language is also demonstrated by neologisms, newly coined words, which occur all throughout history and society. When people hear a word for the first time, they often ask, “Is that a word?” If they ask a linguist, the answer is likely to be, “It is now.” If the novel word is formed according to the morphological and phonological rules of its language and it is understandable in context, it is a bona fide word, even if it’s not found in a dictionary. Consider the word bling, recently coined to mean ‘flashy jewelry.’ It is phonologically well-formed (in English bl is allowed at the beginning of syllables, and the ng [ŋ] sound is allowed at the end). The word has caught on in the mainstream public and is now a bona fide word. Most of these spontaneous coinings – inspired by a particular context, and often labeled as slang – are not used frequently enough to ever make it into a dictionary, but some coinings do become part of the lexicon (and are included in some updated dictionaries) because they meet a new need. Coining new words is one productive process by which languages change to meet the changing communicative needs of their speakers.

语言的生产力很大程度上源于这样一个事实：语言是围绕一组有限的原则组织的，这些原则系统地限制了声音、词素、单词、短语和句子的组合方式。一种语言的母语人士无意识地“知道”这些原则，并能用它们来产生和解释无限多样的话语。定义和明确这些原则是研究语法能力的语言学家的目标之一。

The productivity of languages derives, in large part, from the fact that they are organized around a finite set of principles which systematically constrain the ways in which sounds, morphemes, words, phrases, and sentences may be combined. A native speaker of a language unconsciously “knows” these principles and can use them to produce and interpret an infinite variety of utterances. Defining and making these principles explicit is one of the goals of linguists studying grammatical competence.

任意性

Arbitrariness

虽然语言生产力源自一套有限的原则，这些原则系统地限制了声音、词素、单词、短语和句子的表达方式，但 综上所述，语言在音义对应方面是任意的。除了少数例外，词语与其含义之间没有原则性或系统的联系。英语中前三个数字是一、二、三，但中文则是一、二、三。这两种语言都没有“正确”的词来表示这些数字或其他任何东西，因为根本没有这样的东西（Bolton，1982：5）。即使是拟声词，那些应该听起来像它们所命名的声音的拟声词，例如表示声音的词语，如叮咚和咔哒，以及各种动物发出的声音，也是任意的，并且因语言而异。例如，在英语中，狗会说bow wow或woof woof，但在印地语中它会说bho: bho:。希腊狗说gav，韩国狗说mung mung。人们通过各自语言的任意“声音过滤器”来感知这些声音，因此，即使是像狗叫这样看似客观的东西，在语言中实际上也是任意表示的。

While productivity in language derives from a finite set of principles which systematically constrain the ways in which sounds, morphemes, words, phrases, and sentences may be combined, language is arbitrary in its sound–meaning correspondence. With few exceptions, words have no principled or systematic connection with what they mean. In English, the first three numbers are one, two, three – but in Chinese they are yi, er, san. Neither language has the “right” word for the numerals or for anything else, because there is no such thing (Bolton, 1982: 5). Even onomatopoetic words that are supposed to sound like the noise they name – for example, words for sounds, like ding-dong and click and the sounds various animals make – are arbitrary and vary from language to language. In English, for example, a dog says bow wow or perhaps woof woof, but in Hindi it says bho: bho:. Greek dogs say gav and Korean dogs say mung mung. People perceive these sounds through the arbitrary “sound filters” of their respective languages, so even something as seemingly objective as a dog’s bark is in fact represented arbitrarily in language.

一种特定语言所使用的语音清单也是任意的。英语只使用 36 种不同的声音（根据如何分析英语语音系统的差异，声音数量会多一些或少一些）。但正如您将在第 1 章中详细了解到的那样，英语中使用的声音并不全部与讲其他语言所需的声音相同，它们的组合方式也不相同。英语的 36 种声音又任意用 26 个字母表示，其中一些字母代表两个或多个声音（例如gin和gimp中的g），而其他声音则有两种或更多不同的拼写方式（考虑center中的c和sender中的s或cup中的c、kelp中的k和quiche中的qu）。单词和声音的排列模式也是任意的。我们非常清楚tax的含义，但任何以英语为母语的人都毫无疑问地知道没有 xat这样的词。在英语中，形容词位于名词之前 - 因此是fat man；在法语中，名词位于形容词之前，因此法语中名词是homme gros。任意性是手语和口语的共同特征。手语中的一些手势是标志性的——它们看起来就像它们的意思——但大多数手势没有给出它们含义的丝毫线索。

The inventory of speech sounds used by a particular language is also arbitrary. English is spoken using only 36 different sounds (a few more or less, depending on how the English sound system is analyzed). But, as you will learn in detail in Chapter 1, the sounds used in English are not all the same as the sounds needed to speak other languages, nor are they put together in the same way. The 36 sounds of English are in turn arbitrarily represented by 26 letters, some of which stand for two or more sounds (like g in gin and in gimp) while other sounds are spelled in two or more different ways (consider c in center and s in sender or c in cup, k in kelp, and qu in quiche). The patterns into which words and sounds are arranged are also arbitrary. We know perfectly well what tax means but any English speaker knows without a doubt that there is no such word as xat. Adjectives go before nouns in English – so it’s fat man; in French nouns go before adjectives, making it homme gros. Arbitrariness is a property of sign languages as well as spoken languages. Some manual signs in sign languages are iconic – they look like what they mean – but most signs give not the slightest clue to their meaning.

重要的是要记住，任意性并不意味着无序。例如，这意味着一种语言使用的声音及其组合原则本质上并不比任何其他语言更好或更差。同样，这意味着一种语言变体（或方言）排列单词的原则本质上并不比另一种语言更好或更差。例如，许多说标准英语变体的非语言学家认为使用两个表示否定（称为否定一致）的单词是“不正确的”，例如我没有看到任何人。但是，其他语言（如意大利语）的标准变体中也使用 否定一致：

It’s important to remember that arbitrariness doesn’t mean randomness. It means that, for example, the sounds that one language uses and the principles by which they are combined are inherently no better or worse than those of any another language. Likewise, it means that the principles of one language variety (or dialect) for arranging words are inherently no better or worse than those of another. For example, many non-linguists who speak the standard variety of English believe that it is “incorrect” to use two words that express negation (referred to as negative concord), as in I didn’t see nobody. However, negative concord is used in the standard variety in other languages such as Italian:

一些非标准意大利语变体也像标准英语一样使用单数否定形式。语言的这种无序性也许是公众最需要学习的语言学课程之一。这意味着没有一种语言——也没有一种特定社会中的语言变体——是“正确”的说话方式，也没有一个群体会说不合语法。

And some nonstandard varieties of Italian use the singular negative just like standard English. This property of abritrariness in language is, perhaps, one of the most needed linguistics lessons for the general public. It means that no one language – and no one language variety in a particular society – is the “correct” way of speaking, and no group speaks ungrammatically.

依赖上下文

Reliance on context

任意性（声音序列与意义之间或短语中词序之间的关联）的一个必然结果是二元性。因为单词one（按音标发音为 [w ʌ n]）的发音并不一定与数字 1 相关，所以相同的声音序列（但拼写为won）也可用于表示完全不同的东西——动词 to win的过去式（Bolton，1982 ：5）。但是，如果相同的声音序列在同一种语言中可以表示不同的概念，那么当我说 [w ʌ n]时，你怎样才能弄清楚我想要表达的意思呢？答案是，你必须依赖于上下文，这个答案既复杂又显而易见。如果我在名词前说 [w ʌ n] ，比如“[w ʌ n] dog”，你的英语语法知识会让你猜出我的意思是one。另一方面，如果我在名词（或代词）后说 [w ʌ n] ，如“Mary [w ʌ n]”，同样的知识也会让您猜出我指的是win的过去时。

A corollary of arbitrariness – of association between sound sequences and meanings or in the order of words in phrases – is duality. Because there is nothing about the pronunciation of the word one (transcribed phonetically – as it sounds – it would be [wʌn]) that necessarily associates it with the numeral 1, that same sequence of sounds (but spelled won) can also be used to mean something entirely different – the past tense of the verb to win (Bolton, 1982: 5). But if the same sequence of sounds can represent different concepts in the same language, how are you able to figure out which meaning I intend when I say [wʌn]? The answer – which is as complex as it is obvious – is that you rely on its context. If I say [wʌn] before a noun, as in “[wʌn] dog,” your knowledge of English grammar will lead you to guess that I mean one. On the other hand, if I say [wʌn] after a noun (or pronoun), as in “Mary [wʌn],” that same knowledge will lead you to guess that I mean the past tense of win.

依赖于上下文是语言的一个重要特性，它不仅能帮助我们理解“one”和“won”等词的含义，还能帮助我们解读整个话语的含义。句子的含义主要取决于说出该句子的上下文。该上下文可能是紧接在它之前的句子，也可能是说出该句子的更广泛的物理或社会环境。如果有人说“One”，那么该话语的含义只有在前面的话语的上下文中才会清楚 — — 例如，“你想要一块糖还是两块？”同样，“这里好冷”可能是抱怨，要求关窗，甚至是赞美（也许是关于冰箱的）。给定代词（如“ she”、“it”、“us”或“them”）指代的人或物可能依赖于前面的句子或直接的物理环境。语言依靠形式（所说的内容）和上下文（何时、何地、由谁说以及对谁说）之间的联系来传达比单词序列所包含的内容更多的信息。

Reliance on context is a crucial property of languages, not just in figuring out the meaning of words like one and won, but in interpreting the meaning of entire utterances. The meaning of a sentence depends crucially on the context in which it is uttered. That context could be the sentence or sentences that immediately precede it, or it could be the broader physical or social circumstances in which the sentence it uttered. If someone says “One,” the meaning of that utterance is only clear in the context of a preceding utterance – for example, “Do you want one lump of sugar or two?” Similarly, “It’s cold in here” could be a complaint, a request to close a window, or even a compliment (about a freezer, perhaps). Who or what a given pronoun (like she, it, us, or them) refers to may rely on prior sentences or the immediate physical environment. Languages rely on the connection between form (what is said) and context (when, where, by whom, and to whom it is said) to communicate much more than is contained in a sequence of words.

变异性

Variability

尽管所有语言都具有一些共同的特征，但语言之间也存在许多差异。人们使用的语言因说话者和说话环境的不同而不同。事实上，多变性是语言最重要且最令人钦佩的特性之一。多变性（也称为差异和多样性）是信息的本质。没有光频率的变化，就不会有视觉；没有声音频率的变化，就不会有语言和音乐。（我们开始意识到，如果没有一定程度的遗传多样性，我们的生态系统就会受到威胁。）语言的多变性使人们能够传达的远不止他们所说的单词和句子的语义内容。语言的多变性是指示性的。说话者使用不同的语言来表明他们的社会身份（地理、社会地位、种族甚至性别），并定义当前的言语情况。

Although all languages share some universal characteristics, languages also differ in many ways. The language that people use varies depending on who’s speaking and the situation in which they’re speaking. In fact, variability is one of the most important – and admirable – properties of language. Variation (also known as difference and diversity) is the essence of information. Without variation in light frequencies, there would be no sight; without variation in sound frequencies, there would be no speech and no music. (And as we are beginning to realize, without a certain minimum level of genetic diversity, our ecosystem is threatened.) Variability in language allows people to communicate far more than the semantic content of the words and sentences they utter. The variability of language is indexical. Speakers vary the language they use to signal their social identities (geographical, social status, ethnicity, and even gender), and also to define the immediate speech situation.

人们通过他们所使用的语言的多样性让世界知道他们是谁。他们只需说几句话就能透露他们的地理和社会地位的出身。人们他们还使用多种语言来表明自己属于一系列重叠的社会群体——男性或女性、青少年或成年人、特定族群的成员。尽管老师竭尽全力改变他们的语言，但他们仍保留着自己的语言，因为在潜意识层面，保持与自己出身的联系比任何改变的理由都重要。

People let the world know who they are by the variety of their language that they use. They reveal their geographical and social status origins after saying just a few words. People also use their variety of language to signal membership in a range of overlapping social groups – as male or female, as a teenager or an adult, as a member of a particular ethnic group. They keep their speech, often despite the best efforts of teachers to change it, because at an unconscious level, maintaining their ties to their origin is more important than any reason to change.

人们还使用语言变化来传达他们说话的情形和目的，以及他们在这些情形中扮演的角色。牧师在布道时使用的语言形式与在教堂礼拜后的社交时间不同，扮演不同的角色（并将不同的角色投射到他所面对的教徒身上）。在工作中，人们对下属和上级说话的方式不同，在咖啡休息期间和在会议上说话的方式也不同。父母对自己的孩子说话的方式与对其他成年人（甚至对别人的孩子）说话的方式也不同。书面使用的语言通常与口语中使用的语言不同，反映和传达了语言产生的不同条件及其各种目的。

People also use language variation to communicate the situation and purpose in which they are talking, as well as the roles they are playing in those situations. A priest uses different forms of language during a sermon than during the social hour after a church service, playing different roles (and projecting different roles on the churchgoers he addresses). At work, people speak differently to subordinates than to superiors, and differently during coffee breaks than in meetings. Parents speak differently to their children than to other adults (or even to other people’s children). The language used in writing typically differs from the language used in speaking, reflecting and communicating the different conditions under which language is produced and its various purposes.

语言社区的文化很大一部分是通过语言变化这一媒介来传达的。适当的语言使用规范有助于说话者构建和协商彼此之间的关系。语言的各种形式和用法的非书面和无意识应用的规则可能因文化环境、社会内部和社会之间，甚至性别而异。当说话者不知不觉地按照不同的文化规范行事时，这会增加误解的风险，但当理解这些差异时，我们看待世界的方式就会丰富起来。

A large part of a speech community’s culture is transacted through the medium of language variation. Norms of appropriate language use help speakers to construct and negotiate their relations to each other. The unwritten and unconsciously applied rules for the various forms and uses of language can vary from one cultural milieu to another, within and between societies, and even between genders. This raises the risk of misunderstanding when speakers unknowingly are behaving according to different cultural norms, but enriches our ways of seeing the world when those differences are understood.

语言变异也是语言变化的机制。每当一个新词被创造出来，语言的词汇就会发生一点变化。它的声音清单及其相互关系会随着时间的推移而变化，有时是由于移民或与其他语言的接触，有时是由于其语言社区内部的创新（见第 9 章）。句子中允许的单词顺序也会发生变化（见第 8 章）。甚至规定性规则也会随着时尚或政策的发展而改变（见 第 11 章）。

Language variation is also the mechanism by which languages change. The lexicon of a language changes just a bit every time a new word is coined. Its inventory of sounds, and their relations to each other, changes over time, sometimes due to migration or contact with another language, sometimes due to innovations from within its speech community (see Chapter 9). The order of words allowed in sentences can change as well (see Chapter 8). Even the prescriptive rules can change with developments in fashion or policy (see Chapter 11).

语言变异的后果之一是，任何一种语言的变体或方言都不可能比其他的更好；每种变体或方言都只是语言变化过程中的一个缩影。语言学家发现，有时从共时性角度（作为一个固定系统）来看待语言在分析上很有用，但语言系统总是会发展成一个新的系统。约翰·麦克沃特（John McWhorter ）（ 1998）反对“纯”标准英语的神话，他写道：

One of the consequences of language variation is that no variety or dialect of a language can be better than any other; each is simply a snapshot in the process of language change. Linguists find it analytically useful sometimes to look at language synchronically (as a fixed system), but it is a system always developing into a new system. John McWhorter (1998), arguing against the myth of a “pure” standard English, wrote:

任何语言都在不断演变成一种新的语言，而我们认为“不准确”和不正确的许多声音、词义和句型正是未来“正确”语言的组成部分……我们所认为的“偏离常态”只不过是从一个人的一生的角度来看待语言变化的样子。

Any language is always and forever on its way to changing into a new one, with many of the sounds, word meanings, and sentence patterns we process as “sloppy” and incorrect being the very things that will constitute the “proper” language of the future … What we perceive as “departures from the norm” are nothing more or less than what language change looks like from the point of view of a single lifetime.

考虑一下法语、意大利语和西班牙语都是从拉丁语发展而来的，曾被认为是拉丁语的“腐败”版本。我们现在称之为标准的英语是社会政治偶然的结果，是从 1300 年代英国权力中心的方言发展而来的。我们或许能够消除对讲英语的人的大量歧视如果更多的人理解每种语言和语言的方言都是连贯的，并且同样有效，那么“非标准”变体系统。

Consider that French, Italian, and Spanish each developed from Latin and were once considered “corrupt” versions of Latin. The variety of English we now call standard is the result of a sociopolitical accident, developing from the dialect of the center of British power in the 1300s. We might be able to eliminate a lot of discrimination against speakers of “nonstandard” varieties if more people understood that each language and dialect of a language is a coherent, and equally valid, system.

描述性方法

The descriptive approach

语言是人类的普遍特征，这一事实意味着所有语言（和语言变体）都是平等的。也就是说，它们都来自相同的基因蓝图，并且都同样是“人类”。语言变体之所以不同，是因为它们随着时间的推移适应了其语言社区的不同需求。每种语言的做法都不同：一些语言明确区分几种动词时态（英语只标记两种）；一些语言将名词组织成许多“性别”类别（英语没有）。每种语言在满足其自身语言社区的交流需求方面都具有同等的“功能性”。但有时当两个或多个语言社区接触时，一个群体会比其他群体拥有更多的权力、地位或经济资源。毫不奇怪，该主导群体的语言变体通常也被认为具有更高的地位，特别是如果使用该语言可以获得更多权力或财富的话。相比之下，较弱群体使用的语言变体通常被视为“不正确”或“不好”的语言。

The fact that language is a universal characteristic of human beings means that all languages (and language varieties) are equal. That is, they all come from the same genetic blueprint, and they all are equally “human.” Language varieties differ because over time they have adapted to the differing needs of their speech communities. Each language does things differently: some languages explicitly distinguish between several verb tenses (English marks only two); some languages organize nouns into many “gender” categories (English does not). Each language is equally “functional” at meeting the communicative needs of its own speech community. But sometimes when two or more speech communities come into contact, one group will have more power, status, or economic resources than the others. Not surprisingly, the language variety of that dominant group is often perceived as having higher status as well, especially if speaking it affords increased access to power or wealth. By comparison, the language varieties spoken by the less powerful groups often are stigmatized as “incorrect” or “bad” language.

语言学家研究语言的方式与天文学家研究宇宙或人类学家研究人类文化系统的方式相同。天文学家“错误地”谈论行星绕恒星运转是荒谬的，而人类学家仅仅因为某种文化与自己的文化不同就宣称其“退化”也是不恰当的。同样，语言学家接受他们发现的语言，而不是试图按照先入为主的标准来规范它。语言学家同样对他们遇到的所有语言形式都充满好奇，无论其使用者的教育程度或社会地位如何。

Linguists approach language in the same way that astronomers approach the study of the universe or that anthropologists approach the study of human cultural systems. It would be ridiculous for astronomers to speak about planets orbiting stars “incorrectly” and inappropriate for anthropologists to declare a culture “degenerate” simply because it differs from their own. Similarly, linguists take language as they find it, rather than attempting to regulate it in the direction of preconceived criteria. Linguists are equally curious about all the forms of language that they encounter, no matter what the education or social standing of their speakers might be.

在大多数社会中，某些语言变体被认为是“正确的”，而其他语言变体则被认为是“不正确的”，这一事实对语言学家来说是一种社会现象，是语言使用的一个方面，需要进行科学探索。由于“正确的”语言 本质上并不比那些被认为是“不正确的”语言变体更好或更差，语言学家热切地寻求发现某些语言变异优于其他语言变异的原因，并研究这些信念的后果。

The fact that, in most societies, some varieties of language are perceived as “correct” while others are considered “incorrect” is, for linguists, a social phenomenon – an aspect of language use to be explored scientifically. Since “correct” language is inherently no better or worse than the varieties that are considered “incorrect,” linguists eagerly seek to discover the reasons for the conviction that some part of language variability is superior to the rest, and to examine the consequences of those beliefs.

这种语言态度（认为一种语言优于其他语言）的一个后果是，人们会认为说“不正确”语言的人在某种程度上是低人一等的，因为他们不会或不能“正确地”说话。他们的“不正确”语言随后被用来为进一步的歧视辩护——例如在教育和就业方面。基于语言使用的歧视基于两个错误的主张：一种语言天生就比其他语言更好，人们可以被教导说“正确”的语言。然而，指导一个人学习和使用口语的自然力量是如此强大，以至于明确教授如何说话几乎无关紧要。如果一个人数学不太好，我们可能有理由认为他或她在学校没有学过数学。阅读和写作也可能如此；如果一个人不会读或写，很可能是他/她出了问题。这个人的学校教育有问题。但口语却不是这样。使用否定一致，如She can't find nothing中，或将knowed写为knew 的人，可能从最熟练的老师那里得到了最好的传统语法规则指导。然而，仅仅了解规则是什么，甚至每天在学校练习几分钟，就能像一米高的松树阻止雪崩一样有效地影响一个人的说话方式。影响口语的最强大特征是它能够标记一个人作为日常生活中最亲近的群体成员的身份。这种力量每次都胜过课堂上的语法教学。

One consequence of these kinds of language attitudes – in which one language variety is considered better than others – is the corollary belief that speakers of “incorrect” varieties are somehow inferior, because they will not or cannot speak “correctly.” Their “incorrect” language is then used to justify further discrimination – in education and in employment, for example. Discrimination on the basis of language use is based on two false propositions: that one variety of language is inherently better than others, and that people can be taught to speak the “correct” variety. However, so powerful are the natural forces that guide how a person learns and uses spoken language that explicit teaching on how to speak is virtually irrelevant. If a person is not very good at mathematics, we are probably justified in assuming that he or she did not learn mathematics in school. The same may well be true of reading and writing; if someone cannot read or write, it is likely that something went wrong with that person’s schooling. But the same is not true with spoken language. A person who uses negative concord, as in She can’t find nothing, or says knowed for knew may have received the best instruction in the rules of traditional grammar from the most skilled teachers available. However, just knowing what the rules are, or even practicing them for a few minutes a day in school, will be as effective in influencing how someone speaks as a meter-high pine tree would be in stopping an avalanche. The most powerful feature influencing spoken language is its ability to mark a person’s identity as a member of the group closest to him/her in everyday life. This power trumps grammar instruction in classrooms every time.

即使受过良好教育的美式英语使用者也不会说“你为什么这样做？”（这在形式上是正确的）；他们会说“你为什么这样做？”他们也不会说“你今天见到了谁？”；而是说“你今天见到了谁？”出于完全相同的原因，非标准英语使用者会说“我没有”，因为他们知道“我没有”被认为是正确的——无论是哪种情况，使用“正确的语法”都会让说话者听起来自命不凡或势利，并使他/她失去同龄人的认可。使用某种方式的语言会让人觉得你正在与对你最重要的人划清界限，这极大地阻碍了你这样做。

Even the best-educated speakers of American English will not say “For what did you do that?” (which is formally correct); they’ll say “What did you do that for?” Nor will they say “Whom did you see today?”; instead it will be “Who did you see today?” For exactly the same reason, a speaker of nonstandard English will say “I ain’t got none,” knowing that “I don’t have any” is considered correct – in either case, to use “correct grammar” would make the speaker sound posh or snobbish and cost him/her the approval of his/her peers. There is an enormous disincentive to use language in a way that makes it seem that you are separating yourself from the people who are most important to you.

事实上，那些说话方式接近认可方式的人可能不是在学校里学会的。他们只是幸运地来自制定正确言语标准的社会阶层。这个社会阶层还控制着学校——以及学校使用和教授的语言种类。具有讽刺意味的是，当孩子们在学校里学会使用社会认可的口语种类时，他们并不是从老师在课堂上明确教授的内容中学会的，而是从调整自己的言语以适应大厅里、操场上和校外其他孩子的言语中学会的，从而获得他们的赞同。

In fact, people who speak in close to the approved way probably did not learn to do so in school. They are just fortunate to come from the segment of society that sets the standards for correct speech. This segment of society also controls its schools – and the language variety used and taught in its schools. Ironically, when children learn to use the socially approved variety of spoken language in school, it is not from what their teachers explicitly teach in class, but rather from adjusting their speech to match the speech of the other children in the halls, on the playground, and outside of school, and thus gain their approval.

语言学的多样性

The diversity of linguistics

与其他语言学教科书不同，本书的每一章都是由一位在该领域教学和研究的语言学家撰写的。语言学领域就像它所研究的语言现象一样，内容广泛且多样，尽管语言学家们对某些观点有着共同的看法——例如，他们都认同描述性方法，并且认为所有语言变体都具有同等的功能——但他们在分析时所采用的一些假设却有所不同。一些语言学家——尤其是语音学和音系学、形态学、句法学、语义学/语用学和历史语言学领域的语言学家——在不同程度上认为，语言的形式可以脱离其使用来理解。关于这些主题的章节主要讨论语言形式，构成了 20 世纪中叶语言学的核心内容。自那时起，该领域已大大扩展，本书旨在代表这一更广泛的范围。

Unlike other linguistics textbooks, each chapter in this book has been written by a linguist who teaches and does research in that area. The field of linguistics, like the phenomenon of language which it studies, is broad and diverse, and although linguists share some beliefs – in a descriptive approach, and in the functional equality of all language varieties, for example – they differ in some of the assumptions they bring to their analyses. Some linguists – particularly those in the areas of phonetics and phonology, morphology, syntax, semantics/pragmatics, and historical linguistics – assume, to varying degrees, that the forms of language can be understood separately from their use. The chapters on these topics are primarily about language form and constitute what was considered the essential core of linguistics in the mid twentieth century. Since then, the field has expanded considerably, and this book is designed to represent that broader scope.

如今，语言学研究的不仅仅是语言形式及其意义的具体细节，还包括语言的习得方式（包括第一语言和第二语言），以及语言在反映和创造谈话的互动和文化背景方面所发挥的核心作用，如何设计计算机来处理语言，以及语言如何在我们的大脑中呈现。由于该领域扩展的范围很大程度上涉及对实际使用语言的人的深入研究，因此人们对语法能力和交际能力实际上如何区分开产生了一些疑问。不同的语言学家对语法能力和交际能力之间的区别程度有不同的理解，即使是本书的撰稿人也是如此。

Today the field of linguistics studies not just the nuts-and-bolts of forms and their meanings, but also how language is learned (both as a first and second language), how it plays a central role in reflecting and creating the interactive and cultural settings of talk, how computers can be designed to deal with language, and how language is represented in our very brains. Because much of this expanded scope of the field involves intense study of people actually using language, some doubt has arisen about how separate grammatical and communicative competence actually are. The degree of distinction between grammatical and communicative competence is understood differently by different linguists, even among the contributors to this book.

例如，第 3 章“句子的结构”将句法主要描述为形式，而不是用法。 第 4 章“意义”主要涉及语义意义（语法能力的一部分）和语用意义（交际能力的一部分）之间的界限。第 1 章“语言的声音”中也存在类似的对比，涉及语音学（可观察和观察到的发音现象）和发音系统中更抽象的音系方面（语法能力的一部分）之间的界限。

For example, Chapter 3, “The structure of sentences,” presents syntax as essentially about form, not use. Chapter 4, “Meaning,” is largely about the boundary between semantic meaning (part of grammatical competence) and pragmatic meaning (part of communicative competence). A somewhat similar contrast is seen in Chapter 1, “The sounds of language,” involving the boundary between phonetics (the observable and observed phenomena of pronunciation) and the more abstract phonological aspects of pronunciation systems (a part of grammatical competence).

第 8 章“语言变化”和第 12 章“写作”隐含地介绍了语言形式和交际用法，但较少关注它们之间的严格界限。 第 5 章“话语”、第 6 章“儿童语言习得”和第 13 章“第二语言习得”的作者赞同交际能力的首要地位。在第 10 章“语言与文化”中，我们发现了语法和交际能力不可分割性的最强烈表现，因为作者建议用“语言文化”取代“语言”和“文化”这两个术语，以强调语言与其语境的不可分割性。

Chapter 8, “Language change,” and Chapter 12, “Writing,” implicitly present both language form and communicative use, but with less concern for strict boundaries between them. The sympathies of the authors of Chapter 5, “Discourse,” Chapter 6, “Child language acquisition,” and Chapter 13, “Second language acquisition,” are with the primacy of communicative competence. In Chapter 10, “Language and culture,” we find the strongest representation of the inseparability of grammatical and communicative competence, as the author suggests replacing the terms “language” and “culture” with “languaculture” to emphasize the inseparability of language from its context.

我们如何理解这些强调上的差异？也许语言学家就像盲人一样约翰·戈弗雷·萨克斯 (John Godfrey Saxe) 版本的印度盲人摸象传说。这首诗讲述了几个盲人各自触摸大象不同部位并争论他们触摸的是什么东西的故事。触摸象牙的盲人和触摸尾巴的盲人的想法截然不同。也许语言学研究大体如此，每个语言学家都“部分正确”，这取决于语言学家研究语言的哪些方面。有些方法可能走进死胡同，最终可能不得不放弃。或者也许有人会发现一个绝妙的统一见解，如物理学中的广义和狭义相对论和量子场论，它展示了所有这些方法是如何相互关联的。我们对世界和人类行为的理解在所有这些方面都有所进步。在这本书中，我们向读者介绍了语言研究的当前广度和多样性——以及它们之间的矛盾和未解决的问题。

How are we to understand these differences in emphasis? Perhaps linguists are like the blind men in John Godfrey Saxe’s version of the Indian legend about the blind men and the elephant. The poem relates the story of several blind men each touching a different part of the elephant and arguing over what sort of thing they are touching. The blind man touching the tusks of the elephant has a very different idea from the one holding the tail. Perhaps linguistic research is much the same, and each linguist is “partly in the right,” depending on which aspects of language the linguist is studying. Some approaches could be proceeding down blind alleys and may eventually have to be abandoned. Or perhaps someone will discover a brilliant unifying insight, like general and special relativity and quantum field theory in physics, which shows how all these approaches are related. Our understanding of the world and of human behavior has advanced in all these ways. In this book, we introduce our readers to the study of language in its current breadth and diversity of approaches – with their tensions and unresolved issues.

如何阅读这本书

How to approach this book

在结束本介绍时，我们提出了一些关于如何阅读本书的建议。虽然我们希望本书的内容对每个人都有用，但语言是一种广泛而多样的现象，探索如此广阔的领域有很多方法。语言的不同方面（以及语言学）与其他各种学科更相关，因此我们提出了几条穿越不同语言领域的“路线”。

We close this introduction with suggestions for how to approach this book. While we would like to think that all of it will be useful to everyone, language is a broad and diverse phenomenon and there are many ways to explore such a vast terrain. Different aspects of language – and linguistics – are more relevant to various other academic disciplines, and so we propose several “routes” through different territories of language.

• 普通语言学：语言的声音；单词及其组成部分；句子的结构；含义；语言变化；写作；方言变异；话语
• General linguistics: The sounds of language; Words and their parts; The structure of sentences; Meaning; Language change; Writing; Dialect variation; Discourse
• 语言习得、学习和教学：语言的声音；单词及其组成部分；句子的结构；含义；儿童语言习得；第二语言习得；话语；文化中的语言；语言的政治
• Language acquisition, learning, and teaching: The sounds of language; Words and their parts; The structure of sentences; Meaning; Child language acquisition; Second language acquisition; Discourse; Language in culture; The politics of language
• 社会语言学：语言的声音；单词及其组成部分；含义；语言变化；写作；方言变异；话语；文化中的语言；语言的政治
• Sociolinguistics: The sounds of language; Words and their parts; Meaning; Language change; Writing; Dialect variation; Discourse; Language in culture; The politics of language
• 语言处理：语言的声音；单词及其组成部分；句子的结构；含义；大脑和语言；计算语言学；儿童语言习得
• Language processing: The sounds of language; Words and their parts; The structure of sentences; Meaning; The brain and language; Computational linguistics; Child language acquisition
• 语言学和其他领域：儿童语言习得；文化中的语言；计算语言学；语言政治；大脑与语言；写作；方言变异；话语；第二语言习得
• Linguistics and other fields: Child language acquisition; Language in culture; Computational linguistics; The politics of language; The brain and language; Writing; Dialect variation; Discourse; Second language acquisition

多年来，我们一直为语言学专业和非专业的学生教授多节语言学入门课程，起初使用的是其他教科书，最近使用的是本教科书的第一版。最初，我们遵循“传统”结构；从共享的结构基础开始（第 1-4 章：“语言的声音”；“单词及其组成部分”；“句子的结构”；和“含义”），然后允许不同部分的教师根据他们的专业知识和兴趣选择剩余的章节。几年前，我们转向了一种教学大纲，将结构章节与后续章节更直接地整合在一起。“语言的声音”是接下来是“方言变异”，其中许多语音学和音系学的概念都以学生亲身经历的语境来举例说明。“词语及其组成部分”之后是“语言变化”；“句子结构”之后是“儿童语言习得”（或“语言与大脑”）；“意义”之后是“话语”（或“语言与文化”）。这种平衡的主题顺序使学生能够将语言的结构方面与语言的各种表现形式联系起来，并为每个部分的教师留出了选择其他几章的空间。

We have taught a multi-section introductory linguistics course for both linguistics majors and non-majors for many years, at first with other textbooks and more recently with the first edition of this textbook. Initially, we followed a “traditional” structure; starting with a shared structural foundation (Chapters 1–4: “The sounds of language”; “Words and their parts”; “The structure of sentences”; and “Meaning”), and then allowing instructors of different sections to choose among the remaining chapters according to to their expertise and interest. Several years ago, we shifted to a syllabus which integrates the structural chapters with the following chapters more directly. “The sounds of language” is followed by “Dialect variation,” where many of the concepts of phonetics and phonology are exemplified in contexts that students have personally experienced. “Words and their parts” is followed by “Language change”; “The structure of sentences” is followed by “Child language acquisition” (or “Language and the brain”); and “Meaning” is followed by the chapter on “Discourse” (or “Language and culture”). This balanced sequence of topics enables students to relate the structural aspects of language to a variety of its manifestations, and leaves room for the instructor of each section to choose a couple of other chapters as well.

本教材的配套网站为www.cambridge.org/fasold。在那里，您可以找到声音文件、进一步的阅读建议、额外的练习和教学材料、链接以及关键图表的电子版。

The companion website for this textbook can be found at www.cambridge.org/fasold. There you will find sound files, further reading suggestions, additional exercises and teaching materials, links, and electronic versions of key figures.

1 语言的声音

1 The sounds of language

---

---

章节预览

Chapter preview

本章讲述的是语音。没有声音，交流依然可以进行——点头、挥手、拍照或绘画。甚至没有声音也可以有语言：听不见的人会使用基于手势的语言。然而，对于我们大多数人来说，大多数时候，传达信息需要将其编码为声音。即使在写作时，我们也使用基于语音的符号（尽管有时不是非常直接）。

This chapter is about the sounds of speech. Without sound, communication can still take place – with a nod or a wave, a photograph, or a drawing. There can even be language without sound: those who cannot hear use languages based on manual signs instead. Yet for most of us most of the time, getting our message across involves encoding it in sounds. Even when we write, we use symbols that are based on speech (though sometimes not very directly).

语音研究可分为语音学和音系学两门学科。语音学将语音作为物理对象进行研究。语音学家会提出以下问题：

The study of the sounds of speech can be divided into the disciplines of phonetics and phonology. Phonetics studies speech sounds as physical objects. Phoneticians ask questions such as:

• 语音是如何产生的？
• How are speech sounds made?
• 身体特征如何使人们的声音有所不同？
• How do physical characteristics make people sound different?
• 语言使用多少种不同的声音？
• How many different sounds do languages use?
• 不同的语言和方言如何通过其所使用的声音来区分？
• How are different languages and dialects distinguished by the sounds they use?
• 声音如何在空气中传播？
• How does sound travel through the air?
• 它是如何被耳朵感知的？
• How is it registered by the ears?
• 我们如何衡量言语？
• How can we measure speech?

音系学研究语言如何将声音组织成不同的模式。音系学家会提出以下问题：

Phonology studies how languages organize sounds into different patterns. Phonologists ask questions such as:

• 语言如何组织声音来区分不同的词语？
• How do languages organize sounds to distinguish different words?
• 语言如何限制或约束声音序列？
• How do languages restrict, or constrain, sequences of sounds?
• 如果出现不符合限制的序列，声音会发生哪些变化（交替）？
• What sorts of changes (alternations) do sounds undergo if sequences arise that don’t obey the restrictions?
• 声音如何组织成更大的组成部分（音节、单词、短语）？
• How are sounds organized into larger constituents (syllables, words, phrases)?
• 不同语言和方言中的这些模式有何相同之处？又有何不同之处？
• How are these patterns the same, and how are they different, across languages and dialects?

我们从语音学开始，即研究语音如何产生和感知的学科，然后讨论音系学，即研究语言如何将语音组织成一个有意义的系统。

We begin with phonetics, the study of how speech sounds are made and perceived, and then discuss phonology, the study of how a language organizes those speech sounds into a meaningful system.

目标

Goals

本章的目标是：

The goals of this chapter are to:

• 描述声道的基本解剖结构
• describe the basic anatomy of the vocal tract
• 解释声道结构是如何被控制来发出语音的
• explain how the structures in the vocal tract are controlled to make speech sounds
• 演示如何使用 IPA 音标转录英语单词
• show how to transcribe English words using IPA transcription
• 描述语音超音段方面的基本属性，以及不同语言在使用它们方面有何不同
• describe the basic properties of suprasegmental aspects of speech, and how languages differ in their use of them
• 描述声波的一些物理特性
• describe some of the physical properties of sound waves
• 解释波形、音高轨迹和频谱图的一些基本方面
• interpret some basic aspects of waveforms, pitch tracks, and spectrograms
• 解释音位和音位变体分布
• explain phonemic and allophonic distributions
• 描述一些最常见的音位变化
• describe some of the most common phonological alternations
• 介绍音系理论的一些主要目标
• introduce some of the major goals of phonological theories

您拿起电话，对着话筒说“你好？”。电话那头的声音回应道“嗨！”，随着这个音节，您的大脑中涌现出大量信息。首先，您认出了消息的内容：这是英语中传统的问候语，不同于发音相似的告别（bye！）、疑问（how?或why?）或命令（Fly！）。但您也大概能认出说话人的身份（无论有没有来电显示），并了解他或她的心理状态：兴奋、无聊、高兴或愤怒。如果您无法识别此人，您大概仍能辨别出说话人是男性还是女性，弄清楚他或她是不是英语母语人士，并大概猜出他或她来自哪里。所有这些信息怎么能浓缩成一个单词呢？

You pick up the phone and say “Hello?” into the receiver. The voice on the line responds “Hi!” and with this one syllable your brain is flooded with information. You recognize, first, the content of the message: a conventional greeting in English, distinct from the similar-sounding farewell (bye!), question (how? or why?), or command (Fly!). But you also probably recognize the identity of the person speaking, with or without caller ID, and get a sense of his or her mental state: excited, bored, happy, or angry. If you can’t identify the person, you can still probably tell if the speaker is male or female, figure out whether he or she is a native speaker of English, and make a pretty good guess about where he or she is from. How can all this information be packed into a single word?

事实上，语音构成了一个复杂的多级代码，每个能说会道的人都掌握着它的钥匙。意义被映射到说话者大脑的声音序列中，大脑发出命令来移动声道，从而在空气中或电话线上产生特征振动，影响听者的耳朵和听觉神经，听者的大脑解码信息，找出信号中与信息相对应的方面以及与声音特征相对应的方面传递信息，重新创建声音序列，从而重新创建说话者想要表达的意思。本章将介绍这种编码和解码过程在语音的组织、产生和感知过程中是如何发生的。

Speech sounds in fact constitute an elaborate multilevel code, to which every competent speaker holds the key. Meanings are mapped into sound sequences in the brain of the speaker, which sends commands for vocal tract movements, which produce characteristic vibrations in the air or on the phone line, which impact on the ear and auditory nerve of the listener, whose brain decodes the message, factoring out the aspects of the signal that correspond to the message and those that correspond to the characteristics of the messenger, recreating the sound sequence, and thus the meaning, that the speaker intends. This chapter is about how this process of encoding and decoding takes place in the organization, production, and perception of speech sounds.

发音语音学

Articulatory phonetics

语音学工具

The tools of phonetics

语音学家面临的最大障碍之一是他们无法看到他们正在研究的对象。你看不到舌头在嘴里移动；你看不到声波在空气中传播；你看不到内耳液体的振动。然而，自古以来，语音学家就充分利用了他们所获得的信息，通过仔细聆听、测量、建模和记谱。此外，在过去几十年中，人们开发了更复杂的设备——例如磁共振成像 (MRI)、超声波和数字声学分析等设备。图 1.1显示了这些设备所看到的声道的一些图片。

One of the biggest obstacles phoneticians face is that they can’t see the objects they are studying. You can’t see the tongue as it’s moving around inside someone’s mouth; you can’t see the sound waves traveling through the air; you can’t see the vibration of the fluid in the inner ear. Since ancient times, however, phoneticians have made the best use of the information they had access to, employing careful listening, measuring, modeling, and notation. In addition, more sophisticated devices have been developed within the past decades – devices such as Magnetic Resonance Imaging (MRI), sonography, and digital acoustic analysis. Figure 1.1 shows some pictures of the vocal tract as seen by these devices.

通过这些辅助工具，我们对人类如何发出和听到语音有了哪些了解？

With these aids, what have we learned about how humans make and hear speech sounds?

声道

The vocal tract

基本上，声音就是振动的空气。说话意味着使用声道（肺、气管、喉、嘴和鼻）使空气流动和振动，然后以不同的方式塑造这种运动。图 1.2显示了声道上部的示意图。

Basically, sound is vibrating air. Speaking means using your vocal tract (lungs, trachea, larynx, mouth, and nose) to get air moving and vibrating, and then shaping that movement in different ways. Figure 1.2 shows a diagram of the upper parts of the vocal tract.

大多数语音都是通过肺部的空气发出的；因此，说话从呼吸开始。要开始说话，你需要拉低你的横膈膜，是将胸腔与胃分开的一块大肌肉。它使肺部扩张，从而吸入空气。然后横膈膜放松，肋骨周围的肌肉收缩，慢慢地挤压肺部，迫使空气排出并进入气管。

Most speech sounds are made with air exiting the lungs; therefore, speech begins with breath. To begin to speak, you pull down your diaphragm, the big muscle that separates your chest cavity from your stomach. This enlarges the lungs, which draws air in. Then the diaphragm relaxes and the muscles around the ribs contract, slowly squeezing the lungs and forcing the air out and up the windpipe, or trachea.

气管顶部有一小盒软骨，称为喉（“喉结”）。在喉部内部，两层软组织称为声带（有时称为“声带”），横跨气管顶部。如果声带保持在正确的位置且张力适当，从气管流出的空气会使声带非常快速地（每秒约 200 次）张开和闭合。你可以感觉到这种张开和闭合运动，即喉咙的振动。找到你的喉部（你应该能感觉到喉咙前部的喉结隆起），然后哼唱一段曲调。附着在喉部软骨上的肌肉使你能够调节声带的张力，从而调节振动频率并提高或降低音调。振动越快，音调越高。其他肌肉也使你可以将声带拉开，以免发生振动。

At the top of the trachea is a little box of cartilage, called the larynx (the “Adam’s apple”). Inside the larynx, two folds of soft tissue, called the vocal folds (sometimes called “vocal cords”), lie across the top of the trachea. If the vocal folds are held in the correct position with the correct tension, the air flowing out of the trachea causes them to flap open and closed very quickly (around 200 times per second). You can feel this opening and closing motion as vibration in your throat. Find your larynx (you should be able to feel the bump of the Adam’s apple at the front of your throat), and then hum a tune. Muscles attached to the cartilages of the larynx allow you to adjust the tension of the folds, thus adjusting the rate of vibration and raising or lowering the pitch. The faster the vibration, the higher the pitch of the voice. Other muscles also allow you to draw the folds apart so that no vibration occurs.

会厌位于喉部正上方，舌根处。会厌是一种肌肉结构，吞咽时会向下折叠在喉部上方，以防止食物在进入胃部之前进入肺部。喉部位于喉咙下方的风险会导致口腔后部出现开放区域，咽喉。咽喉使舌头可以自由地前后移动。其他哺乳动物，包括非人类灵长类动物，喉部位于口腔后部高处，与鼻腔相连。由于没有咽喉，黑猩猩永远学不会说话。（这就是为什么试图教灵长类动物用语言交流的科学家改用基于手势的语言。）

Just above the larynx, at the base of the tongue, is the epiglottis. The epiglottis is a muscular structure that folds down over the larynx when you swallow to prevent food from going down into the lungs before it enters the passage to the stomach. The payoff for the risk of a larynx located low in the throat is an open area at the back of the mouth, the pharynx. The pharynx allows the tongue freedom for front and back movement. Other mammals, including nonhuman primates, have the larynx high up at the back of the mouth, connected to the nasal passages. Because they have no pharynx, chimps could never learn to talk. (This is why scientists who try to teach primates to communicate with language use gesture-based languages instead.)

口腔内部有许多不同的结构——主动发声器官和被动发声器官——当空气通过声道时，我们用它来塑造语音。主动发声器官向被动发声器官移动，以收缩和塑造从肺部流出的空气。主动发声器官包括嘴唇，可以张开或闭合、撅起或张开，以及舌头。我们通常看到的是小小的粉红色舌尖，但实际上舌头是一大块相互连接的肌肉，充满口腔底部。虽然舌头没有骨头或软骨，但舌头的不同部分可以相当独立地移动。舌前部（包括舌尖和舌叶，从舌尖向后延伸几厘米）、舌体（舌头的主要部分，也称为舌背）和舌根（舌头的最低部分，位于咽喉后部）被认为是独立的活动发声器官。

Inside the mouth itself, there are many different structures – active articulators and passive articulators – that we use to shape speech sounds as the air passes through the vocal tract. The active articulators move toward the passive articulators in order to constrict and shape the air that is moving out from the lungs. Active articulators include the lips, which can be opened or closed, pursed or spread, and the tongue. What we usually see of the tongue is the small, pink tip, but it is actually a large mass of interconnected muscles that fills the floor of the mouth. Although the tongue has no bones or cartilage, different parts of the tongue can move fairly independently. The tongue front (including the tongue tip and the tongue blade, which extends a few centimeters back from the tip), the tongue body (the main mass of the tongue, also known as the dorsum), and the tongue root (the lowest part of the tongue, back in the pharynx), are considered separate active articulators.

被动发声器官位于声道的顶部。从上牙后面开始，用舌头沿着嘴的顶部移动。你首先会碰到牙槽嵴，牙齿后面的骨质隆起。牙槽后区呈拱形从牙槽嵴向硬腭，即口腔的顶部。如果你把舌头往嘴里卷得很远，你可以感觉到硬腭的骨结构让位于更软的组织，也就是所谓的软腭，或软腭。软腭是一种肌肉结构，它调节软腭口，即位于口腔后部的开口，连接口腔和鼻子。当软腭降低时，例如在呼吸时以及在发出 [m] 和 [n] 等声音时，软腭口打开，空气在鼻子和肺部之间自由流动。（将声音符号写在方括号内是一种语音惯例。有关语音书写的更多信息，请参阅以下部分。）当软腭升高时，例如在发出大多数语音时，鼻子的开口会关闭，所有气流都会通过口腔。软腭的最末端是悬雍垂，当你张大嘴巴说“啊。”

The passive articulators lie along the top of the vocal tract. Run your tongue along the top of your mouth beginning behind your upper teeth. You will first encounter the alveolar ridge, the bony rise just behind your teeth. The postalveolar region arches from the alveolar ridge toward the hard palate, the roof of the mouth. If you curl your tongue very far back in your mouth, you can feel that the bony structure of the hard palate gives way to softer tissue, which is known as the soft palate, or velum. The velum is a muscular structure that regulates the velar port, the opening in the back of the mouth that connects the mouth and nose. When the velum is lowered, as it is for breathing and for some sounds such as [m] and [n], the port is open and air flows freely between the nose and lungs. (It’s a phonetic convention to write the symbols for sounds within square brackets. See the following sections for more on phonetic writing.) When the velum is raised, as it is for most speech sounds, the opening to the nose is closed off and all the airstream is directed through the mouth. At the very end of the velum is the uvula, the little pink pendulum you can see hanging down in the back of your mouth when you open wide and say “ah.”

关节

Articulation

说话涉及以不同方式使用声道结构来控制和塑造流动的空气。我们可以想象说话者通过“选择”使用哪些主动和被动发音器官以及如何进行不同的收缩来产生正确的组合。这些选择不是有意识的；它们是经过长期练习而自动产生的，就像走路或伸手的肌肉动作一样。当你伸手去拿一杯咖啡时，你不会对自己说：“好的，现在收缩肱三头肌，放松肱二头肌”等。相反，你会不假思索地选择一个目标，比如“手指握拳”，然后你长期练习的动作就会执行目标导向的动作。说话也是一样。说话的动作是有目标导向的手势。每个声音都由一组发音目标组成，这些目标将使声道处于正确的位置，从而发出你想要发出的声音。像“发出 [m]”这样的总体目标可以分解为一组子部分：“闭上嘴唇”、“打开软腭”、“使声带振动”。这些子程序可以以不同的方式重新组合以发出不同的声音，就像一组乐高积木可以根据它们的组合方式建造一座城堡或一艘船。

Speaking involves using the structures of the vocal tract in different ways to control and shape moving air. We can think of the speaker producing the right combinations by making “choices” about which active and passive articulators to use and about how different constrictions will be made. These choices are not conscious; they are automated from long practice, just like the muscular routines of walking or reaching. When you reach for a cup of coffee, you don’t say to yourself, “OK, now contract the tricep, relax the bicep,” etc. Instead, without much conscious thought you select a goal, such as “fingers to mug,” and your long-practiced routines of motion execute the goal-directed movement. Speech works the same way. The movements of speech are goal-directed gestures. Each sound is comprised of a set of articulatory goals that will get the vocal tract in the right positions to make the sound you wish to make. An overall goal like “make an [m]” can be broken down into a set of subcomponents: “Close the lips,” “open the velum,” “make the vocal folds vibrate.” These subroutines can be recombined in different ways to make different sounds, like a set of Lego blocks that can build a castle or a boat depending on the way they’re put together.

第一个选择是气流机制：扬声器首先如何让空气流动？通常的选择是肺音呼气– 即空气从肺部呼出。世界上大多数语言使用的大多数声音都是肺音呼气。但是，也可以通过其他方式使空气流动，例如上下移动喉部，或者用舌头顶住上颚弹出小气囊，就像咔哒声一样。（咔哒声包括我们在英语中写的tsk tsk或tut tut这样的表达，但在南非的一些语言中，这些声音被纳入到语音流中，成为英语中的常规辅音 [p] 或 [t]。）本章的其余部分仅讨论肺部呼气的声音。

The first choice is that of airstream mechanism: how will the speaker get the air moving in the first place? The usual choice is pulmonic egressive – that is, air moving out from the lungs. Most sounds used by most of the world’s languages are pulmonic egressive. However, it is also possible to get air moving in other ways, such as by moving the larynx up or down, or by popping little pockets of air made with the tongue against the roof of the mouth, as in clicks. (Clicks include the expression we write in English as tsk tsk or tut tut, but in some languages of southern Africa sounds such as these are incorporated into the stream of speech as regular consonants like [p] or [t] in English.) The rest of this chapter discusses only sounds that are pulmonic egressive.

第二个选择是如何处理声带。声带振动发出的声音是浊音；声带不振动发出的声音是清音。如果你把手指放在喉部并发出持续的 [z] 音，你应该能够感觉到振动或浊音。如果你切换到 [s]，即清音，振动就会停止。对于某些音，比如pop中的首字母 [p] ，声带会分开得足够远足够长，让额外的“一股气流”在 [p] 结尾时从嘴里出来。这叫做送气。当你发pop或pill时，如果你将指尖放在嘴唇前一英寸左右，你就能感觉到额外的气流被释放。送气可以用上标h表示：[p ^h ]。

The second choice is what to do with the vocal folds. Sounds produced with vocal fold vibration are voiced; sounds produced without vocal fold vibration are voiceless. If you place your finger on your larynx and produce a sustained [z], you should be able to feel the vibration, or voicing. If you switch to [s], a voiceless sound, the vibration ceases. For some sounds, as in the initial [p] in pop, the vocal folds are held apart far enough and long enough to allow an extra “puff of air” to exit the mouth at the end of the [p]. This is called aspiration. You can feel the extra release of air if you hold your fingertips an inch or so in front of your lips as you say pop or pill. Aspiration can be indicated by a superscripted h: [p^h].

除了决定如何处理喉部之外，说话者还必须决定软腭是否张开。如果软腭张开，空气就会流入鼻子，那么这个声音就是鼻音（如 [m]）。如果软腭闭合，那么这个声音就是口音。

Besides deciding what to do with the larynx, the speaker must decide whether the velum will be open or not. If the velum is open, so that air flows into the nose, the sound is nasal (like [m]). If the velum is closed, the sound is oral.

最后，说话者必须决定使用哪个主动发声器官来做出收缩（嘴唇、舌前、舌体、舌根），以及在哪里做出收缩（发音部位)以及将进行何种收缩 (发音方式)。以下各节将讨论各种发音部位；我们首先讨论各种发音方式。

Finally, the speaker must decide which active articulator will be used to make a constriction (lips, tongue front, tongue body, tongue root), where the constriction will be made (the place of articulation), and what sort of constriction will be made (the manner of articulation). The various places of articulation are discussed in following sections; we turn first to the various manners of articulation.

表达方式

Manners of articulation

发音方式包括：塞音、摩擦音、塞擦音、近音和元音。

The manners of articulation include: stop, fricative, affricate, approximant, and vowel.

如果主动和被动发声器官结合在一起，形成一个完全闭合的状态，这样气流从口腔中完全被切断，发声的方式就是塞音。英语中的 [p]、[t] 和 [k] 音是塞音。非常缓慢地说出单词poppa，注意在单词中间的 [p] 时，嘴巴完全闭合，没有空气从嘴里出来。您甚至可能感觉到嘴唇后面有压力，因为空气继续从肺部流出，无处可去。当嘴唇张开时，这种压力会随着轻微的砰砰声或爆裂声而释放。[m] 音是鼻塞音。即使软腭张开，空气自由地从鼻子流出，这样您就可以在发出 [m] 音的同时哼曲调，但发音方式仍然是塞音，因为嘴唇完全闭合，就像 [p] 一样。（试着在哼曲调时捏住鼻子一会儿，看看会发生什么。）

If the active and passive articulators are brought together to make a complete closure, so that airflow out of the mouth is completely cut off, the manner of articulation is a stop. The sounds [p], [t], and [k] in English are stops. Say the word poppa very slowly, and note that there is complete silence, with no air exiting the mouth, while the lips are closed for [p] in the middle of the word. You may even feel pressure building up behind the lips, as air continues flowing from the lungs and has nowhere to go. This pressure is released with a slight pop, or burst, when the lips are opened. The sound [m] is a nasal stop. Even though the velum is open and air flows freely out of the nose, so that you can hum a tune while producing an [m] sound, the manner of articulation is still a stop, because the lips are completely closed, as they were for [p]. (Try pinching your nose closed for a moment while you’re humming a tune, and see what happens.)

如果发音器官靠得很近但没有完全闭合，那么在它们之间强制流动的气流就会变得湍急和嘈杂，发音方式就是摩擦音。[s]、[z]、[f] 和 [v] 是摩擦音。塞擦音将一系列塞音和擦音组合成一个音。英语中通常写为 ch 的是塞擦音。试着尽可能慢地achoo发音a和oo之间的舌头运动。首先，将舌头前部闭合在齿槽脊处或齿槽脊后方，然后降低舌尖，让空气通过稍靠后、位于舌叶和齿槽后区域之间的狭窄收缩处排出。

If the articulators are brought close together but not closed completely, so that the stream of air that is forced between them becomes turbulent and noisy, the manner of articulation is a fricative. The sounds [s], [z], [f], and [v] are fricatives. Affricates combine a sequence of stop plus fricative in a single sound. The sound usually written ch in English is an affricate. Try saying the word achoo as slowly as possible, paying attention to the movement of the tongue between the a and oo sounds. You first make a closure with the tongue front at or just behind the alveolar ridge, and then lower the tongue tip to let the air out through a narrow constriction slightly further back, between the tongue blade and postalveolar region.

如果主动发音器官移动以缩小声道，但又不至于产生摩擦音，那么发音方式就是近音。滑音，例如单词yell和well开头的音，是近音，就像英语中的 [l] 和 [r] 一样。世界上各种语言的llaterals，因为空气从舌头两侧流出。试着拉长单词lateral。现在，不要移动舌头，深吸一口气。你会感觉到空气从舌头两侧流动。r音被称为rhotics。世界各国语言的rhotics 音种类繁多，包括舌头快速敲击齿龈的音、舌头因气流而振动的颤音，以及美式英语 [r] 的非常奇怪的形状，其中舌体高高地聚拢，舌尖可能向上或向后卷曲。（非英语母语人士很难发出这个音，这并不奇怪。）元音是最开放的发音方式。不同的元音是通过上下、前后移动舌体以及嘴唇的圆润或张开而发出的。然而，在发所有元音时，声道都相对开放，空气可以自由流出。口塞音、摩擦音和塞擦音一起形成一类称为阻塞音，因为它们通过阻碍声道中的气流来产生声音，当封闭被释放时会发出一阵声音，当空气通过狭窄的收缩时会发出嘶嘶声。鼻塞音、近音和元音（任何不是阻塞音的音）形成一类声音，称为响音。它们不是通过阻碍气流而是通过让空气产生共振来发出可听见的声音。响音几乎总是有声的。声带的振动会导致声道内的空气振动。如果振动足够强，就会产生可听见的声音，就像铃声一样。不同的声道​​形状（我们通过移动活跃的发声器官来控制）会产生不同的振动模式，我们听到的是不同的声音（更多内容见下文）。通过在开放的声道中移动大量空气，可以产生无声的响音。像苗语和缅甸语使用清鼻音。仔细听，你会听到英语单词play中的 [l]也是无声。

If the active articulator moves to narrow the vocal tract, but not so much that fricative noise is created, the manner of articulation is an approximant. Glides, such as the sounds at the beginning of the words yell and well, are approximants, as are [l] and [r] in English. The l-sounds of the languages of the world are called laterals, because air flows out over the sides of the tongue. Try drawing out the initial sound in the word lateral. Now, without moving your tongue, take a deep breath in and out. You’ll feel the air moving over the sides of the tongue. The r-sounds are called rhotics. The rhotic sounds of the languages of the world are quite varied, including quick taps of the tongue against the alveolar ridge, trills in which the tongue is set into vibration by air flowing over it, and the very odd shape of the American English [r], in which the body of the tongue is bunched up high and the tongue tip may be raised or curled backwards. (It is no surprise that non-native speakers of English have trouble with this sound.) Vowels are the most open manner of articulation. Different vowel sounds are made by moving the tongue body up or down, front or back, and by rounding or spreading the lips. During all vowel sounds, however, the vocal tract is relatively wide open, and air flows out freely. Oral stops, fricatives, and affricates together form a class of sounds called obstruents, because they make noise by obstructing the airflow in the vocal tract, causing a burst of sound as a closure is released or a hissing sound as the air passes through a narrow constriction. Nasal stops, approximants, and vowels (anything that’s not an obstruent) form a class of sounds called sonorants. They make audible sounds not by obstructing the airflow, but by letting the air resonate. Sonorant sounds are almost always voiced. The vibration of the vocal folds causes the air inside the vocal tract to vibrate. If the vibration is strong enough, it produces an audible sound, like the ringing of a bell. Different vocal tract shapes (which we control by moving the active articulators) produce different patterns of vibration, which we hear as different sounds (more on this below). It is possible to produce voiceless sonorants, by moving a large volume of air through the open vocal tract. Languages like Hmong and Burmese use voiceless nasals. Listen carefully, and you’ll hear that the [l] in an English word like play is also voiceless.

书写声音：转录

Writing sounds: transcription

在讨论英语和其他语言中使用的不同发音位置之前，我们必须考虑如何写下不同的声音。描述性短语，如“ yell单词开头的声音”或“ achoo中间的声音”，都很麻烦。我们需要一个语音字母表。用语音字​​母表写下声音称为音标。

Before we discuss the different places of articulation used in English and other languages, we have to consider how to write down different sounds. Descriptive phrases like “the sound at the beginning of the word yell” or “in the middle of achoo” are cumbersome. We need a phonetic alphabet. Writing down sounds using a phonetic alphabet is called phonetic transcription.

1888 年，国际语音协会（总部设在巴黎）致力于解决如何准确描述其成员在努力描述世界上所有的语言。他们根据以下两个原则 发布了新字母表——国际音标( IPA ) 的符号：

In 1888, the International Phonetic Association (based in Paris) tackled the problem of how to precisely describe any sound the members might encounter in their efforts to describe all the languages of the world. They published symbols for a new alphabet, the International Phonetic Alphabet (IPA), based on two principles:

考虑一下英语拼写是如何不满足这两个原则的。我们当然没有字母来表示南非的喀拉喀斯声，也没有喉咙深处发出的声音。阿拉伯语。一个声音/一个符号的对应关系经常被破坏。字母c有多少种发音方式？（考虑每个恶性循环。）字母x代表两个声音的序列，[k] 后跟 [s]（box与locks押韵），而字母sh序列代表单个声音（舌叶抵住后齿龈区发出的摩擦音）。

Consider how English spelling falls short on these two principles. We certainly don’t have letters for the clicks of southern Africa or the sounds made in the back of the throat in Arabic. And the one sound/one symbol correspondence is constantly violated. How many ways can you pronounce the letter c? (Consider each vicious circle.) The letter x stands for a sequence of two sounds, [k] followed by [s] (box rhymes with locks), while the sequence of letters sh stands for a single sound (a fricative made with the tongue blade against the postalveolar region).

图 1.4显示了本章讨论的所有声音的 IPA 符号，以及更多内容。

Figure 1.4 shows the IPA symbols for all the sounds discussed in this chapter, and more.

在 IPA 辅音中图表中，发音部位写在顶部，发音方式写在侧面，因此每个单元格都表示特定方式和部位的组合。灰色阴影单元格是物理上不可能的组合。如果单元格中有两个符号，则左侧的符号为无声符号，右侧的符号为有声符号。

In an IPA consonant chart, place of articulation is written across the top, and manner of articulation is written down the side, so that each cell indicates a combination of a specific manner and place. Cells that are shaded gray are physically impossible combinations. If there are two symbols in a cell, the one on the left is voiceless, the one on the right is voiced.

学习 IPA 音标的一个好起点是表 1.1中列出的英语辅音符号，以及出现这些声音的示例单词。虽然不是全部，但许多符号都是熟悉的。当您学习不同位置和发音方式的定义时，您应该能够将它们结合起来，找出 IPA 表中不熟悉符号的发音。

A good place to start learning IPA transcription is with the symbols for the consonants of English, which are given in Table 1.1, along with example words in which the sounds occur. Many, though not all, of the symbols will be familiar. As you learn the definitions of the different places and manners of articulation, you should be able to combine them to figure out the pronunciations of unfamiliar symbols in the IPA chart.

当你查看表 1.1中的示例单词时，你会注意到并非所有单元格都已填满。其中一些空白是偶然的：例如，英语恰好没有单词nen ，尽管它有类似的单词nine和net。其他空白是系统性的：例如，英语中没有以 [h] 结尾的单词。（许多单词以字母 h结尾，但我们关心的是声音，而不是拼写。）系统性空白将在后面讨论以下。

As you look through the example words in Table 1.1, you’ll notice that not all of the cells are filled. Some of these gaps are accidental: English just doesn’t happen to have a word nen, for example, though it has similar words like nine and net. Other gaps are systematic: no word in English ends with the sound [h], for example. (Plenty of words end with the letter h, but we are concerned with sound, not spelling.) Systematic gaps will be discussed further below.

辅音

Consonants

我们将按照主动发音器官和发音部位来组织对辅音的讨论。一般来说，每个发音器官可以移动到多个发音部位，如表 1.2所示。

We will organize our discussion of the consonants by active articulator and place of articulation. Generally, each articulator can move to more than one place of articulation, as shown in Table 1.2.

下唇可以在两个不同的地方收缩。如果下唇和上唇合在一起，声音是双唇音。[p]、[b] 和 [m] 是双唇音。请注意，[p] 是清音，而 [b] 和 [m] 是浊音。或者，下唇可以与上牙接触，产生唇齿音。[f] 和 [v] 是唇齿音。日语有双唇音擦音（IPA [ ϕ ] 而不是 [f]）。此音与日语中futon的发音一样，是通过撅起嘴唇吹气发出的，与吹灭蜡烛的动作相同。

The lower lip can make constrictions at two different places. If the lower and upper lip come together, the sound is bilabial. The sounds [p], [b], and [m] are bilabials. Note that [p] is voiceless and [b] and [m] are voiced. Alternatively, the lower lip can make contact with the upper teeth to produce a labiodental sound. [f] and [v] are labiodentals. Japanese has a bilabial fricative (IPA [ϕ] instead of [f]). This sound, as in the native Japanese pronunciation of futon, is made by blowing through pursed lips, the same motion as blowing out a candle.

下唇收缩的部位相当有限。舌前部是活动发音器官中最灵活的，至少可以移动到四个不同的发音部位。舌尖向前移动到上齿，发出“thin”和“”的音。这些齿擦音写作 [θ] (清音) 和 [ð] (浊音)。

The lower lip is rather limited in the places at which it can make a constriction. The tongue front is the most versatile of the active articulators, moving to at least four different places of articulation. The tongue tip moves forward to the upper teeth for the sounds at the beginning of thin and then. These dental fricatives are written [θ] (voiceless) and [ð] (voiced).

英语中 [t]、[d]、[n] 和 [l] 的发音都是舌尖位于齿龈处，齿龈发音部位。摩擦音 [s] 和 [z] 也是齿龈发音。对于这些摩擦音，舌头在齿龈脊下形成一个狭窄的凹槽，就像一个喷口，将一股气流喷向牙齿，产生高音调的嘶嘶声。虽然这些摩擦音的发音部位是齿龈，但前牙对于产生正确的高音调嘶嘶声是必不可少的，正如所有孩子在失去乳牙时发现的那样。

The English sounds [t], [d], [n], and [l] are made with the tongue tip at the alveolar ridge, the alveolar place of articulation. The fricatives [s] and [z] are also alveolar. For these fricatives the tongue forms a narrow groove under the alveolar ridge like a spout that shoots a stream of air against the teeth, producing a high-pitched hissing sound. Though the place of articulation for these fricatives is alveolar, the front teeth are necessary to create the proper high-pitched hiss, as all children discover when they lose their baby teeth.

擦音 [ ʃ ] 和 [ ʒ ]（如pressure和measure中间的音）的发音位置更靠后，舌叶在 后齿龈发音处收缩。（有趣的是，[ ʒ ] 在英语中不出现在首位，除非明显借用了 法语，如流派。）church和judge中的塞擦音也是后齿龈音。这些音的 IPA 符号为 [ ʧ ] 和 [ ʤ ]：两个符号用于表示塞音加摩擦音，用连字符连接。（在其他转录系统中在语言学书籍中，[ ʃ ]、[ ʒ ]、[ ʧ ] 和 [ ʤ ]常用带音节的符号书写：[ š ]、[ ž ]、[ č ]、[ ǰ ]。）英语没有后齿龈塞音或鼻音，但其他语言也有鼻音（例如法语agneau '羔羊'，西班牙语año ‘年’，都发音为 [a ɲ o]，但重音不同之处）。

The fricatives [ʃ] and [ʒ] (as in the middle of pressure and measure) are made further back, with the blade of the tongue making a constriction at the postalveolar place of articulation. (Interestingly, [ʒ] doesn’t occur in initial position in English, except in obvious borrowings from French, such as genre.) The affricates in church and judge are also postalveolar. The IPA symbols for these sounds are [ʧ] and [ʤ]: two symbols are used for the combination of stop plus fricative, linked by a ligature. (In other transcription systems commonly used in linguistics books, [ʃ], [ʒ], [ʧ], and [ʤ] are written with hatchecks: [š], [ž], [č], [ǰ].) English doesn’t have any postalveolar stops or nasals, though other languages do (for example, French agneau ‘lamb’, and Spanish año ‘year,’ both pronounced [aɲo], though with a stress difference).

通常用舌叶来形成齿龈后缩窄。但是，舌尖也可以向后卷曲，使该区域缩窄。如果舌尖向后卷曲，则该音称为回音。（这些声音的 IPA 符号在符号下方有一个小钩，让人联想到舌头的卷曲。）对于一些（但不是全部）美国语使用者来说，[r] 是回音近音。你能确定你自己的舌尖在单词road？英语中没有其他回音，尽管其他语言，尤其是印地语和其他印度语言，都有一套完整的回音塞音、擦音和鼻音。英语中印度口音的一个明显标志是用回音塞音代替英语的齿龈音。

Usually the blade of the tongue is used to make a postalveolar constriction. It is also possible, however, for the tip of the tongue to curl back to make a constriction in this area. If the tip of the tongue curls back, the sound is called retroflex. (The IPA symbols for these sounds have a little hook under the symbol, recalling the curling back of the tongue.) For some (but not all) American speakers, [r] is a retroflex approximant. Can you determine whether your own tongue tip curls back in a word like road? There are no other retroflex sounds in English, though other languages, notably Hindi and other languages of India, have a full set of retroflex stops, fricatives, and nasals. A telltale sign of an Indian accent in English is substituting retroflex stops for English alveolars.

英语单词you和yacht的开头发音是腭音（确切地说是腭滑音）。整个中段舌体（包括舌叶和舌体）被向上推，以缩小舌头和硬腭之间的空间。腭滑音的 IPA 符号是 [j]。（想想斯堪的纳维亚语ja。）英语没有其他腭音，但它们并不难发音。从 [j]（如you）开始，然后将其清音。结果是清音腭擦音，例如用于德语单词，例如ich [iç]，意思是“我”。

The sound at the beginning of the English words you and yacht is palatal (a palatal glide, to be exact). The whole middle section of the tongue, including blade and body, is pushed straight up to narrow the space between the tongue and hard palate. The IPA symbol for a palatal glide is [j]. (Think Scandinavian ja.) English doesn’t have any other palatal sounds, but they’re not hard to make. Start with [j] (as in you), and then make it voiceless. The result is a voiceless palatal fricative, used for example in German words such as ich [iç], meaning ‘I’.

继续向声道后方移动，下一个发音部位是软腭音，此时舌体向上移动，紧贴口腔后部高处的软腭。英语中的 [k] 和 [g] 是软腭塞音。在英语中，字母ng的序列，例如在song或ring的末尾，通常表示软腭鼻音。在单词song中，不是发出一连串的腭鼻音后跟软腭塞音（ng），而是在与 [k] 或 [g] 相同的位置发出一个鼻音。（感受一下当你说单词king时你的舌头移动有多小。）软腭鼻音的 IPA 符号是 [ ŋ ]。与 [ ʒ ] 一样，英语只在单词末尾（song）或中间（songs ）使用 [ ŋ ] ，从不在单词开头使用，不过通过练习，你可以学会发音如下单词泰语 [ ŋ a:] 'tusk' 或澳大利亚名字，如Ngaio 。Bach 这个名字末尾的德语发音是清音软腭擦音，[x]。要发这个音，先从 [k] 开始，然后稍微放松收缩，让少量湍流通过。

Moving further back in the vocal tract, the next place of articulation is velar, in which the tongue body moves up to make constriction against the velum, high in the back of the mouth. The English sounds [k] and [g] are velar stops. In English, the sequence of letters ng, as at the end of song or ring, usually indicate a velar nasal. In the word song, you don’t make a sequence of alveolar nasal followed by velar stop (n-g), but a single nasal sound at the same place as [k] or [g]. (Feel how little your tongue moves when you say the word king.) The IPA symbol for a velar nasal stop is [ŋ]. As with [ʒ], English uses [ŋ] only at the end of words (song) or in the middle (singer), never at the beginning, although with practice you can learn to pronounce words like Thai [ŋa:] ‘tusk’ or Australian names like Ngaio. The native German sound at the end of the name Bach is a voiceless velar fricative, [x]. To make this sound, begin with a [k], then loosen the constriction slightly, letting a little turbulent airflow pass through.

舌体还可以进一步收缩至小舌塞音的发音部位。要发出小舌塞音，先发 [k] 或 [g]，然后将舌头向后移动几厘米。小舌塞音在许多美洲原住民语言中很常见，在阿拉伯语和希伯来语。卡塔尔国名的阿拉伯语发音以小舌塞音开始。喉咙深处也可以收缩，舌根向后移向咽壁。这些浊音和清音 在阿拉伯语和希伯来语中也发现 咽擦音。

The tongue body can also make constrictions further back, at the uvular place of articulation. To make a uvular stop, begin with a [k] or [g], then move the tongue a few centimeters back. Uvular stops are common in many Native American languages, as well as in Arabic and Hebrew. The native Arabic pronunciation of the country name Qatar begins with a uvular stop. Constrictions can also be made deep in the throat, with the tongue root moving back toward the pharyngeal wall. These voiced and voiceless pharyngeal fricatives are also found in Arabic and Hebrew.

最后，辅音可以用喉部作为唯一的发音器官来发音。[h] 音由气流通过张开的声带发出的声音组成，可以被认为是喉部 摩擦音。也可以将声带紧紧关闭，阻止喉部气流，喉塞音（IPA [ʔ]）。这是英语的中间表达“啊哦”。如果你慢慢地发这个音，你可以感觉到喉咙收缩。在其他语言中，比如夏威夷语中，喉塞音用作常规辅音。在夏威夷这个词中，撇号代表喉塞音。喉塞音还代表某些英语单词中的某些 /t/ 音，但具体哪些单词受到影响取决于方言：对于许多美国人来说，button和mitten等单词有喉塞音；在伦敦部分地区，kitty和pretty有喉塞音停止。

Finally, consonants can be made with the larynx as the only articulator. The sound [h] consists of the noise of air rushing through the open vocal folds, and may be considered a laryngeal fricative. It is also possible to close the vocal folds up tight, stopping the airflow at the larynx, a glottal stop (IPA [ʔ]). This is the sound in the middle of the English expression uh-oh. If you pronounce this slowly, you can feel the constriction in the larynx. In other languages, like Hawai‘ian, the glottal stop is used as a regular consonant. In the word Hawai‘i the apostrophe stands for a glottal stop. A glottal stop also stands in for certain /t/ sounds in some English words, though exactly which words are affected depends on the dialect: for many Americans, words like button and mitten have glottal stops; in parts of London kitty and pretty have glottal stops.

还有一个英语辅音需要讨论：滑音 [w]，如wear中的滑音。这个音结合了软腭发音部位声道的收窄 嘴唇圆润。因此，这是双发音，即唇腭滑音。虽然在发音部位的不同位置进行双发音并不难（考虑到主动发音器官的独立性），但耳朵很难区分，因此除唇腭滑音之外的双发音很少见。

One English consonant remains to be discussed: the glide [w], as in wear. This sound combines a narrowing of the vocal tract at the velar place of articulation with rounding of the lips. It is thus a double articulation, a labiovelar glide. While double articulations at various places of articulation are not hard to make (given the independence of the active articulators), they can be hard for the ear to distinguish, so double articulations other than labiovelars are rare.

总结一下，有 11 个常见的发音部位——双唇音、唇齿音、齿音、齿龈音、后齿龈音、卷舌音、腭音、软腭音、小舌音、咽音和喉音——尽管没有一种语言会用所有的发音部位来发辅音。 关节。

In summary, there are eleven common places of articulation – bilabial, labiodental, dental, alveolar, postalveolar, retroflex, palatal, velar, uvular, pharyngeal, and laryngeal – though no single language makes consonants using all of the places of articulation.

元音

Vowels

元音比辅音更难描述。根据定义，元音有一个开放的声道，因此舌头实际上不会在任何特定位置接触声道的上表面，而发音部位这个术语实际上并不是元音的发音方式并不恰当。相反，不同的元音是根据舌体和嘴唇的运动方式来描述的。语言学家根据舌体的高度、舌体是向嘴的前部还是后部聚拢以及嘴唇是否圆润来对元音进行分类。

Vowels are harder to describe than consonants. By definition, vowels have an open vocal tract, so the tongue doesn’t actually touch the upper surface of the vocal tract at any particular place and the term place of articulation isn’t really appropriate. Instead, different vowels are described in terms of the ways in which the tongue body and lips move. Linguists classify vowels by the height of the tongue body, whether it is bunched toward the front or back of the mouth, and whether the lips are rounded.

如果描述一般的元音系统是一项艰巨的任务，那么描述英语的元音就更是如此。一个原因是因为英语的元音数量很多。一种语言最常见的元音数量是五个。尽管英语作者只使用五个字母来编码他们的元音（旧系统的遗留），但英语使用了十几种不同的元音。另一个原因是，不同方言的元音数量和元音质量差异很大，远大于辅音的差异。例如，珀斯人发音的mate听起来很像（尽管不完全像）巴尔的摩人发音的might 。亚特兰大人发音的my听起来类似于波士顿人发音的mar ，以及西雅图人发音的ma。对于美国东海岸的大多数说话者来说，caught和cot这两个词有两个不同的元音；但对于西海岸的大多数说话者来说，这两个词的发音相同。

If describing vowel systems in general is a difficult task, describing the vowels of English is even more so. One reason is because there are a lot of them. The most common number of vowels for a language to have is five. Though English writers use just five letters to encode their vowels (relics of an older system), the English language uses more than a dozen different vowel sounds. Another reason is because the exact number of vowels and exact vowel quality differ from dialect to dialect, much more so than for the consonants. For example, the word mate as pronounced by a speaker from Perth sounds a lot (though not exactly) like the word might as pronounced by a speaker from Baltimore. The word my spoken by a native of Atlanta sounds similar to mar as pronounced in Boston, and ma as pronounced in Seattle. For most speakers on the East coast of the United States, the words caught and cot have two different vowel sounds; but for most speakers on the West coast, the two words are pronounced the same.

表 1.3列出了“通用美式英语”（GA）元音的 IPA 符号和示例词，GA 即美国中部地区或多或少使用的英语（“或多或少”是因为每种英语方言都有特定的地域特征，而纯正的“通用美式英语”的理想使用者并不存在）。在英国，BBC 播音员和学校教师所追求的“标准”方言被称为RP（即“标准发音”）；通常是这种方言，而不是美式英语，更接近世界各地使用的多种英语版本，例如新加坡、南非和印度英语。（其他英语变体，如澳大利亚英语和苏格兰英语，有自己独特的系统。）本节列出了 GA 和 RP 元音之间的一些主要发音差异；请参阅本章末尾列出的参考资料以及第 9 章，了解有关英语的许多地区和国际变体的更多信息。

Table 1.3 gives the IPA symbols, with example words, for the vowels of “General American” (GA) English – that is, English as it is more or less spoken in the central United States (“more or less” since every dialect of English has particular regional characteristics, and the ideal speaker of pure “General American” doesn’t exist). In England, the “standard” dialect to which BBC announcers and school teachers aspire is known as RP (for “Received Pronunciation”); it is usually this dialect, rather than American, which is closer to the many versions of English spoken around the world, such as Singaporean, South African, and Indian English. (Other varieties, such as Australian and Scottish, have their own unique systems.) Some of the major pronunciation differences between GA and RP vowels are noted in this section; see the references listed at the end of this chapter, as well as Chapter 9, for more information on the many regional and international varieties of English.

图 1.5大致基于发音时舌头最高点的位置，绘制了（GA）英语元音的相对位置。元音空间在顶部比在底部大，因为靠近上颚的地方有更多的空间供舌头移动。（敏锐的读者会注意到图 1.4和1.5中元音符号的位置存在一些差异，特别是低元音和中央元音。这是因为 IPA 提供的符号比任何一种语言所需的符号都多得多。如果语言学家主要关心的是区分一种语言的单词所需的一组声音的一般描述（就像我们在这里一样），他/她通常会从 IPA 提供的符号中选择更熟悉的符号。）

Figure 1.5 charts the positions of (GA) English vowels relative to each other, based roughly on the position of the highest point of the tongue during that vowel. The vowel space is larger at the top than at the bottom, because there is more room for tongue movement closer to the palate. (The astute reader will notice some differences in exactly where the vowel symbols are placed in Figures 1.4 and 1.5, particularly for low and central vowels. This is because the IPA provides many more symbols than are needed in any one language. If a linguist is concerned primarily with a general description of the set of sounds that are needed to distinguish the words of a language (as we are here), he/she will usually choose the more familiar symbols from among those the IPA makes available.)

这我们用来区分不同元音的术语指的是发元音时舌头的最高点。高元音[i, ɪ , ɨ , u, ʊ ] 时舌体向上移动，发低元音[æ, a] 时，舌头在中间，发中元音[e,ɛ, o,ɔ,ʌ,ə] 时，舌头在嘴里向前移动。前元音[i,ɪ, e,ɛ, æ] 和后元音后元音[u,ʊ, o,ɔ, a]。元音 [ɨ,ʌ,ə] 位于中央。元音还因以下方面而不同：唇形圆润。在普通美式英语中，后元音 [u, ʊ , o, ɔ ] 是圆音，其他元音都是不圆音。在普通英语中， body等单词中的元音也是圆音：这种低、后、圆音的符号为 [ ɒ ]。此外，英语将其元音分为两组，紧张和松弛。紧元音 [i, e, o, u] 比松弛元音 [ ɪ,ɛ,ɔ,ʊ更长、稍高，发音时舌根僵硬程度更大。紧/松弛区别并不适用于低元音。这些描述性术语可以组合起来选出一个特定的元音：[ɪ] 是高、前、松弛、不圆音；[o] 是中、后、紧、圆音。

The terms we use to classify different vowels refer to the highest point of the tongue during the vowel. The tongue body moves up for the high vowels [i, ɪ, ɨ, u, ʊ], down for the low vowels [æ, a], and stays in the middle for the mid vowels [e, ɛ, o, ɔ, ʌ, ə]. The tongue moves forward in the mouth for the front vowels [i, ɪ, e, ɛ, æ] and backward for the back vowels [u, ʊ, o, ɔ, a]. The vowels [ɨ, ʌ, ə] are central. Vowels also differ with respect to lip rounding. In General American English, the back vowels [u, ʊ, o, ɔ] are round, all other vowels are unround. In RP, the vowel in words like body is also round: the symbol for this low, back, round vowel is [ɒ]. In addition, English divides its vowels into two sets, tense and lax. The tense vowels [i, e, o, u] are longer, slightly higher, and produced with greater stiffening of the tongue root than their lax counterparts [ɪ, ɛ, ɔ, ʊ]. The tense/lax distinction doesn’t really apply to low vowels. These descriptive terms can be combined to pick out a specific vowel: [ɪ] is high, front, lax, unround; [o] is mid, back, tense, round.

只出现在短而非重读音节中的一组元音使用不同的符号。（参见下文关于重读的讨论。）中元音 [ ə ] 称为中元音，出现在aboutRosa’s的第二个音节。高音、中央的 [ɨroses的第二个音节中。

Distinct symbols are used for the set of vowels that occur only in short, unstressed syllables. (See the discussion of stress below.) The mid-central vowel [ə], called schwa, is heard in the first syllable of about and the second syllable of Rosa’s. The high, central [ɨ] occurs in the second syllable of roses.

超音段

Suprasegmentals

到目前为止，我们已经了解了单个声音，但说话涉及将声音串联成更大的单元。影响大于单个音段的声音延伸的语音方面称为超音段。语音的超音段方面包括长度、音调、语调、音节结构和重音。由于语音的超音段方面涉及将声音组织成更大的单元，因此超音段的研究跨越了语音学（研究语音作为物理对象）和音系学（研究语言如何将声音组织成不同的模式）的领域。

Thus far, we have learned about individual sounds, but speaking involves stringing sounds together into larger units. Aspects of speech that influence stretches of sound larger than a single segment are called suprasegmentals. Suprasegmental aspects of speech include length, tone, intonation, syllable structure, and stress. Because suprasegmental aspects of speech involve the organization of sounds into larger units, the study of suprasegmentals straddles the domains of phonetics (the study of speech sounds as physical objects) and phonology (the study of how languages organize sounds into different patterns).

---

---

---

---

长度

Length

许多因素都会影响发声某个音节所需的时间。有时，元音长度的差异是元音发音方式不同的无意结果。发低元音时嘴巴必须张大，因此发声所需的时间比发高元音时所需的时间长，而发高元音时几乎不需要移动。然而，在某些语言中，两个音节可能仅在长度上有所不同：长音节和短音节完全相同，只是前者（有意）保持的时间更长，即多了一个“节拍”。这些长音节可以用双符号（[aa]、[pp]）或在通常的符号后加一个冒号（[a:]、[p:]）来表示。（有时使用水平线或长音符号来表示长音节。）例如，日语对元音和辅音的长度都有区别。在东京，你要小心点 [bi:.u]“一杯啤酒”，而不是 [bi.u]“一栋建筑”，并询问前往某条 [tori]“街道”的方向，而不是 [to:.i]“门”或 [to:.i:]“鸟”。我们将保留“长元音”一词来指代日语中的区别，其中长/短元音对的所有内容都相同，除了长度之外。

Many factors influence how long it takes to articulate a given segment. Sometimes differences in vowel length are unintentional results of how different vowels are articulated. Low vowels, for which the mouth has to open wide, take longer to articulate than high vowels, for which little movement is necessary. In some languages, however, two segments may differ in length alone: the long segment and short counterpart are exactly the same, except that the former is (intentionally) held for a longer period of time, an extra “beat.” These long segments may be written with a double symbol ([aa], [pp]) or with a colon after the usual symbol ([a:], [p:]). (Sometimes a horizontal bar, or macron, is used to indicate a long segment.) For example, Japanese makes length distinctions in both vowels and consonants. In Tokyo, you want to be careful to order [bi:.u] ‘a beer’ rather than [bi.u] ‘a building,’ and to ask directions to a certain [tori] ‘street,’ rather than a [to:.i] ‘gate’ or [to:.i:] ‘bird.’ We will reserve the term “long vowel” to refer to distinctions such as those in Japanese, where everything about the long/short vowel pair is the same, except for length.

长辅音被称为双辅音。当两个单词连在一起时，英语可以产生长辅音——比较bookcase[bʊkkes] 和book ace[bʊkes] 或top part[tappa.t] 和top art[tapa.t]——但我们不会区分单词中的长辅音和短辅音。当写出双辅音时，例如supper和super，它们实际上告诉我们元音的质量，而不是辅音的长度。

Long consonants are known as geminates. English can create long consonants when two words come together – compare bookcase [bʊkkes] to book ace [bʊkes] or top part [tappa.t] to top art [tapa.t] – but we don’t distinguish long and short consonants within words. When double consonants are written, for example in supper vs. super, they actually tell us about the quality of the vowel, not the length of the consonant.

---

---

声调和语调

Tone and intonation

声音的音调包含很多信息。它可以告诉你说话者是男是女、是大是小、是老是少。高音调可以告诉你一个人很害怕；低音调则表明他或她很生气。然而，这类信息实际上不是语言信息，而是生理或情感信息。声调和语调这两个术语指的是音调的语言用法。声调是指使用音调在单词层面传达含义；语调是指使用音调在句子或话语层面传达含义。

The pitch of the voice carries a lot of information. It can tell you whether the speaker is a male or female, a large person or small, old or young. High pitch can tell you that a person is frightened; low pitch that he or she is angry. This sort of information isn’t really linguistic, however, but physical or emotional. The terms tone and intonation refer to linguistic uses of pitch. Tone refers to the use of pitch to convey meaning at the word level; intonation refers to the use of pitch to convey meaning at the sentence or discourse level.

语调可以区分不同类型的句子或将注意力集中在某个特定的单词上。例如，尝试大声（生动地）朗读以下句子：

Intonation distinguishes different kinds of sentences or focuses attention on a particular word. For example, try reading the following sentences out loud (and dramatically):

在说单词cat时，您的音调会朝不同的方向变化。说第一个cat 时，音调上升，表示疑问。说第二个 cat 时，音调下降，表示陈述或确认。说第三个cat时，音调的下降-上升模式更为复杂，表示怀疑。（在印刷术上，我们分别用问号、句号和斜体表示这些不同的“读法”。）在每种情况下，序列 [k ^h æt] 都指的是同一个对象，即猫科动物。音调差异只表示在当前对话中对猫科动物的提及所起的作用：询问有关猫的信息、提供信息或表达对所提供信息的怀疑。所有语言都会在一定程度上使用语调，尽管不同语言的模式和含义有所不同。

The pitch of your voice moves in different directions on the word cat. On the first cat, pitch goes up, indicating a question. On the second, pitch falls, indicating a statement or confirmation. On the third cat, a more complicated fall–rise pattern indicates incredulity. (Typographically, we indicate these different “readings” with a question mark, period, and italics, respectively.) In each case, the sequence [k^hæt] refers to the same object, a feline. The pitch differences indicate only the role that the reference to the feline is playing in the current conversation: asking for information about the cat, providing it, or expressing disbelief regarding the information offered. All languages use intonation to some extent, though the patterns and meanings differ across languages.

除了语调，大多数语言还使用音调来区分不同的单词。在英语中，无论你用上升音调还是下降音调说 [k ^{h æt]，这个词仍然是指猫科动物。在}泰语中，如果你用上升音调说 [k ^{h a:]，它的意思是“腿”；但如果你用下降音调说，它的意思是“值”。（泰语中实际上有五种对比鲜明的音调模式：高、低、中、下降和上升。）这些词对英语使用者来说就像}cat和cut一样不同。这种使用音调来区分不同单词的方式称为声调。

In addition to intonation, most languages also use pitch to distinguish different words. In English, whether you say [k^hæt] with a rising pitch or falling pitch, the word still refers to a feline. In Thai, if you say [k^ha:] with rising pitch, it means ‘leg’; but if you say it with falling pitch, it means ‘value.’ (There are actually five contrasting pitch patterns in Thai: high, low, mid, falling, and rising.) These words are as different as cat and cut to an English speaker. This use of pitch, to distinguish different words, is known as tone.

虽然声调的概念对于英语使用者来说似乎很奇怪，但世界上大多数语言都是有声调的。欧洲主要语言及其近亲是例外，因为它们没有声调。语气。

Although the idea of tones seems very strange to English speakers, the majority of the world’s languages are tonal. The major European languages and their relatives are exceptional in not having tone.

音节结构

Syllable structure

Appalachicola？Massachusetts？Antidisestablishmentarianism ？这些单词有多少个音节？（分别为六个、四个和十一个。）说英语的人很容易就能数出一个单词的音节数，但语言学家很难定义什么是音节。

How many syllables are in the word Appalachicola? Massachusetts? Antidisestablishmentarianism? (six, four, and eleven, respectively.) English speakers have little trouble counting the number of syllables in a word, but linguists have a harder time defining what a syllable is.

一个初步的答案可能是“一个元音和它周围的辅音”。我们在英语中遇到的大多数音节，比如pin、print甚至sprints等词，都符合这个定义。但是，完全有可能有一个没有元音的音节。我们都同意hidden有两个音节，即使发音为 [h ɪ dn]，两个辅音之间也没有元音。此外，将音节定义为“一个元音和它周围的辅音”并不能解释为什么某些辅音序列是允许的，而其他的则不可以。序列 [pr ɪ nt] 在英语中可以作为音节，但序列 [rp ɪ tn] 在任何语言中都不能作为单个音节。为什么 [pr ɪ nt] 是一个好音节，而 [rp ɪ tn] 不是？

One preliminary answer might be “a vowel and its surrounding consonants.” Most of the syllables we encounter in English, in words like pin, print, or even sprints, fit this definition. However, it’s perfectly possible to have a syllable without a vowel. We would all agree that hidden has two syllables, even if pronounced [hɪdn], with no vowel between the two consonants. Also, defining a syllable as “a vowel and the consonants around it” doesn’t explain why some sequences of consonants are allowed and others are not. The sequence [prɪnt] is acceptable as a syllable in English, but the sequence [rpɪtn] is not acceptable as a single syllable in any language. Why is [prɪnt] a good syllable when [rpɪtn] is not?

最好的答案（尽管并不完美）在于响度的概念。响度可以定义为声道的相对开放程度，它直接对应于声音的相对响度。最响亮的声音是低元音；嘴巴张得很大，声音自由流出。最不响亮的声音是清音塞音；嘴巴完全闭合，完全不发出声音。其他声音介于这两个极端之间。

The best answer (though not perfect) lies in the concept of sonority. Sonority can be defined as relative openness of the vocal tract, which corresponds directly to the relative loudness of a sound. The most sonorous sounds are the low vowels; the mouth is wide open, and the sound flows freely out. The least sonorous sounds are the voiceless stops; the mouth is completely shut, and no sound is made at all. Other sounds range between these two extremes.

语音流由声音的峰和谷组成。语言通常不会选择长串的辅音或长串的元音。相反，我们交替 更响亮的声音和不太响亮的声音：每个声音都在另一个声音的背景下显得更加突出。因此，音节可以被定义为围绕响度峰值组织声音的一种方式。

The speech stream is organized into peaks and valleys of sonority. Languages generally do not choose long strings of consonants nor long strings of vowels. Rather, we alternate sounds that are more sonorous and less sonorous: each stands out better against the background of the other. A syllable, then, may be defined as a way of organizing sounds around a peak of sonority.

以单音节pin为例。元音 [ ɪ ] 是序列中最响亮的声音，其两侧是响度较低的辅音。因此，只有一个响度峰值和一个音节。音节印记也遵循响度原则。响度从 [p] 开始上升（清塞音 ) 变为 [r] (第一个音节是 [rhotic] 音节，从 [ rhotic ] 音节变为 [ɪ ] 音节，然后从元音变为 [n] 音节，再变为 [t] 音节。只有一个峰值，一个音节。而 [rp ɪ tn] 音节有三个峰值，响度较高的 [r]、[ ɪ ] 和 [n] 被响度最低的 [p] 和 [t] 所打断。因此（如果它可以发音的话）它有三个音节，而不是一个。

Take the simple syllable pin. The vowel [ɪ] is the most sonorous sound in the sequence, flanked by less sonorous consonants. Thus there is a single sonority peak, and a single syllable. The syllable print also follows the principle of sonority. Sonority rises from [p] (voiceless stop) to [r] (rhotic) to [ɪ] (vowel), then falls from vowel to [n] (nasal) to [t] (stop). A single peak, a single syllable. Meanwhile, the sequence [rpɪtn] has three peaks; higher sonority [r], [ɪ], and [n] are interrupted by lowest sonority [p] and [t]. Thus (if it is pronounceable at all) it has three syllables, not one.

音节中最响亮的元素，即音峰本身，称为音核。音核之前的低响度声音称为音首；音核之后的声音称为音节尾声。核心和尾声共同构成韵。图 1.6显示了print的音节结构树形图。（在语言学中，这种树形图通常用于显示较大单位的组成部分之间的关​​系，因此您将在其他章节中看到更多树形图！）

The most sonorous element of a syllable, the peak itself, is called the nucleus. Lower sonority sounds preceding the nucleus are called the onset; those following the nucleus are called the coda. The nucleus and coda together form the rhyme. A syllable structure tree diagram for the word print is shown in Figure 1.6. (In linguistics, such tree diagrams are often used to show how constituent parts of a larger unit are related, so you’ll see more tree diagrams in other chapters!)

由于元音是最响亮的声音，它们通常构成音节核，但情况并非总是如此。除元音之外的声音也可能形成响度峰值，因此hidden和prism有一个元音但有两个音节。

Since vowels are the most sonorous sounds, they usually constitute syllable nuclei, but that’s not always the case. Sounds other than vowels may form sonority peaks, so that hidden and prism have one vowel but two syllables.

因此，响度似乎捕捉到了我们对音节结构的大部分直觉，并解释了世界上各种语言中可能出现的音节。但响度并不能解释一切。有些英语单词明显违反了响度原则——例如sprints和sixties。语言学家不确定如何处理这样的单词。可能像复数 [s] 和序数 [ θ ] 这样的结尾根本不是音节的一部分；而是作为“附录”附加在音节的末尾。单词。

Sonority thus seems to capture most of our intuitions about syllable structure and explains a lot about possible syllables in the languages of the world. But sonority doesn’t account for everything. There are some English words that clearly violate the principle of sonority – sprints and sixths, for example. Linguists aren’t sure exactly how to deal with words like this. It may be that endings like plural [s] and ordinal [θ] are not really part of the syllable at all; rather they’re tacked on in an “appendix” to the end of the word.

压力

Stress

语言重音是音节之间的突出关系：某些音节比周围的音节更长、更响、音调更高或发音更清晰。就像我们通常可以数出单词中的音节一样，我们通常可以选出最突出的音节：音系学、语音学、句法学。

Linguistic stress is a prominence relation between syllables: certain syllables are longer, louder, higher-pitched, or more clearly articulated than those around them. Just as we can generally count the syllables in a word, we can generally pick out the syllable that’s most prominent: phoNOlogy, phoNEtics, SYNtax.

英语中至少有三种不同的重音级别。以 Alabama这个词为例。第三个音节 [bæ] 最为突出，承担着整个音节的主要或主要重音。单词，但其他三个音节的重音并不相等。第一个音节具有完整的元音质量 [æ]，尽管它不如第三个音节那么长或响亮，但第二和第四个音节短而弱，舌头不会远离其中心位置。我们说第一个音节有次要重音，而第二和第四个音节完全没有重音。

There are at least three different levels of stress in English. Consider the word Alabama. The third syllable [bæ] is the most prominent, and bears the main or primary stress of the word, but the other three syllables do not receive equal stress. The first syllable has a full vowel quality [æ], though it is not quite as long or loud as the third, but the second and fourth syllables are short and weak, with the tongue not moving far from its central position. We say that the first syllable has secondary stress, while the second and fourth are completely unstressed.

有些语言不使用重音全部。例如，在日语中，单词中的所有音节有时发音相同。这给说英语的人带来了不少麻烦。例如，1996 年日本举办冬季奥运会时，英语评论员在决定如何说出主办城市的名字时遇到了很大麻烦：长野？NaGAno？NagaNO？事实上，这个词的所有音节都同样重读（或同样非重读），这对说英语的人来说是一项艰巨的任务。相反，说日语的人经常在减少英语的非重读元音方面遇到很大困难，这在英语人士听来过于谨慎。

Some languages do not use stress at all. In Japanese, for example, all the syllables in a word are sometimes pronounced with equal prominence. This gives quite a bit of trouble to English speakers. In 1996, for example, when the Winter Olympics were held in Japan, English commentators had a great deal of trouble deciding how to say the name of the host city: NAgano? NaGAno? NagaNO? In fact, the word is pronounced with all syllables equally stressed (or equally unstressed), a difficult task for an English speaker. Conversely, Japanese speakers often have a great deal of trouble with reducing the unstressed vowels of English, sounding overly careful to English ears.

因为重音是一种突出关系，所以重音和非重音音节往往会在单词中交替出现：例如ApalachiCOla。为了使音节听起来突出，最好将其周围安排一些非突出音节。在英语中，重音有时甚至会移动以适应交替模式。在像sixteen这样的单词中，重音通常落在第二个音节上（你几岁了？SixTEEN）。但是，如果将这个单词放在以强重音音节开头的单词旁边，重音可能会向后移动以保持交替模式（你在这里工作了多久？SIXTEN YEARS）。语言学家将重音音节和相邻非重音音节的组合称为音步。选择不同类型的音步（da-DUM 与 DA-dum）不仅在诗歌中很重要，而且是定义语言展开的语音节奏的一部分。

Because stress is a prominence relation, stressed and unstressed syllables tend to alternate across the word: ApalachiCOla, for instance. In order for a syllable to be heard as prominent, it helps if it’s surrounded by non-prominent syllables. In English, stress sometimes even moves around in order to accommodate an alternating pattern. In a word like sixteen, stress usually falls on the second syllable (How old are you? SixTEEN.). But put the word next to one that begins with a strongly stressed syllable, and stress may shift back in order to maintain an alternating pattern (How long have you worked here? SIXteen YEARS.) Linguists refer to a grouping of a stressed syllable and adjacent unstressed syllables as a foot. Choosing different kinds of feet (da-DUM vs. DA-dum) is important not only in poetry, but as part of defining a language’s unfolding rhythm of speech.

重音完全可预测的语言被称为固定应力系统。固定应力可以是交替的，例如Pintupi 是一种澳大利亚语言，重音总是落在第一个音节上，然后落在之后的每个音节上：[KUra ululimpatjura]“第一个（是）我们的亲戚。”其他固定重音系统可能会在单词中挑选出一个特别突出的音节。在波斯语中，重音总是在第一个音节上，土耳其语总是在最后一个音节上，总是在倒数第二个音节进行抛光。

Languages in which stress is completely predictable are called fixed stress systems. Fixed stress may be alternating, as in Pintupi, an Australian language where stress always falls on the first syllable and then on every other syllable after that: [KUra ululimpatjura] ‘the first one (who is) our relation.’ Other fixed stress systems may pick out just one syllable in the word as especially prominent. In Farsi, stress is always on the initial syllable, in Turkish always on the last syllable, in Polish always on the second-to-last syllable.

在其他语言中，重音是不可预测的：你只需要记住学习单词时哪个音节要重读，就像你必须记住它是以 [b] 还是 [p] 开头一样。这被称为词汇重音。例如，俄语有词汇重音系统：如果重音放在第一个音节上，[duxi] 的意思是“烈酒”；重音放在第二个音节上，则意思是“香水”。

In other languages, stress is unpredictable: you just have to memorize which syllable gets stressed when you learn the word, in the same way you have to memorize whether it begins with [b] or [p]. This is called lexical stress. For example, Russian has a lexical stress system: if stress is placed on the first syllable, [duxi] means ‘spirits’; stressed on the second syllable, it means ‘perfume.’

第三类系统是聚合重音，重音模式取决于单词的词性——例如，名词或动词。英语系统大多是聚合重音，有一些不可预测的方面。一般来说，英语动词和形容词遵循一套规则，名词遵循另一套规则。因此，英语中有只在重音上不同的单词对，其中一个单词是名词，另一个是动词：我们用 reJECT 表示 REject，用 reCORD 表示 REcord，用 convert 表示 CONVERT，用 inSULT 表示侮辱等等。

The third type of system is paradigmatic stress, in which the stress patterns depend on what part of speech a word is – for example, a noun or a verb. The English system is mostly paradigmatic, with some unpredictable aspects. Generally, English verbs and adjectives follow one set of rules, nouns another. Thus English has pairs of words that differ only in stress, where one word is a noun and the other a verb: we reJECT the REject, reCORD the REcord, conVERT the CONvert, inSULT with an INsult, etc.

声学语音学

Acoustic phonetics

到目前为止，我们已经讨论了发音，即语音是如何在口腔内发出的。但口腔内发生的事情只是这个过程的一部分。为了理解人们如何使用声音进行交流，我们还必须了解发音器官如何将空气运动转化为声音，声音经过嘴唇后会发生什么，它如何在空气中传播，以及它如何影响听众的耳朵和大脑（有时还有麦克风、录音机和电脑）。语言学对声音的研究属于声学语音学。

Thus far, we have talked about articulation, how speech sounds are made inside the mouth. But what happens inside the mouth is only part of the process. In order to understand how people use sound to communicate, we must also understand how the articulators turn air movements into sound, what happens to sound after it passes through the lips, how it travels through the air, and how it impacts on the ears and brain (and sometimes the microphones, recorders, and computers) of those who listen. These aspects of the linguistic study of sound fall under the heading of acoustic phonetics.

声波

Sound waves

语音是由流动的空气产生的。发音就是让空气以可以听到的方式流动——振动、爆裂或嗖嗖。在这个简短的介绍中，我们将重点介绍浊音（尤其是元音）的振动模式。但语音非常复杂，所以从思考更简单的东西开始会有所帮助——音叉。

Speech sounds are caused by moving air. Articulation is all about getting air to move in ways that can be heard – vibrating, popping, or swishing. In this short introduction, we will focus on the patterns of vibration in voiced sounds, particularly vowels. But speech sounds are very complex, so it helps to start by thinking about something much simpler – a tuning fork.

将音叉敲击在桌子或其他硬物上，叉齿就会振动。根据音叉的具体形状和大小，叉齿会以特定的速率或频率振动。例如，调到“管弦乐 A”的音叉将以每秒 440 次 (cps) 的频率振动，其中一个周期等于叉齿末端来回运动一次。（“每秒周期”通常称为“赫兹”——缩写为 Hz——以一位著名物理学家命名，海因里希·赫兹。）

Strike a tuning fork against a table or other hard object, and the tines of the fork will vibrate. Depending on the exact shape and size of the fork, the tines will vibrate at a particular rate, or frequency. A fork tuned to “orchestral A,” for example, will vibrate at a frequency of 440 cycles per second (cps), where a cycle equals one back and forth motion of the end of the tine. (“Cycles per second” are often called “Hertz” – abbreviated Hz – after a famous physicist, Heinrich Hertz.)

不同的物体具有不同的固有振动频率，这决定了声音的音调。小手铃的振动速度更快，发出的声音也更高，而巨大的教堂钟的振动速度更慢，发出的声音也更低。人类可以听到低至约 20 Hz 和高达约 20,000 Hz 的频率。人类不会将比这更快或更慢的振动视为声音，但狗可以听到高达 45,000 Hz 的频率，而蝙蝠和海豚可以听到超过 100,000 Hz 的频率。

Different objects have different inherent frequencies of vibration, which determine the pitch of the sound. A small handbell will vibrate faster and make a higher pitched sound than a huge churchbell, which will vibrate slower and make a lower pitched sound. Human beings can hear frequencies as low as about 20 Hz, and as high as about 20,000 Hz. Humans will not perceive vibrations faster or slower than that as sound, although dogs can hear frequencies up to 45,000 Hz and bats and dolphins over 100,000 Hz.

音叉的振动是如何以声音的形式传到我们耳朵里的？当音叉的末端振动时，它们会使旁边的空气粒子也振动，并进行相同的来回运动。这些移动的空气粒子交替推拉旁边的粒子，以及旁边粒子上的粒子，依此类推，因此振动模式就像池塘里的涟漪一样从音叉向外移动。这些移动的振动模式称为声波。当声波到达我们的耳朵（或麦克风）时，它们会使耳膜（或麦克风中的膜）按照相同的模式振动。在耳朵内部，振动会引发神经冲动，我们的大脑会将其解释为声音。

How does the vibration of the tuning fork make it to our ears as sound? As the ends of the tuning fork vibrate, they set the air particles next to them vibrating as well, following the same back and forth motion. These moving air particles alternately push and pull on the particles next to them, and those on the particles next to them, and so on, so that the pattern of vibration moves outward from the tuning fork like ripples in a pond. These moving patterns of vibration are called sound waves. When the sound waves reach our ears (or a microphone), they set the eardrum (or the membrane in the microphone) vibrating according to the same pattern. Inside the ear, the vibrations set off nerve impulses, which are interpreted by our brains as sound.

简单和复杂的声音

Simple and complex sounds

音叉的运动非常简单（像钟摆一样来回摆动），因此它产生的声波也非常简单。如此简单的振动发出的声音是单一频率（如 440 Hz）的纯音。但声道的振动——以及它发出的声音——则更为复杂。

The motion of a tuning fork is very simple (back and forth, like a pendulum), and thus the sound wave it creates is also very simple. The sound such a simple vibration makes is a pure tone of a single frequency (like 440 Hz). But the vibrations of the vocal tract – and the sounds it makes – are more complicated.

声道就像单簧管。声带是簧片，是声音的来源。 振动。口中的气柱是乐器中的气柱，是振动的过滤器。而说话者当然是音乐家，通过改变气柱的形状来调节产生的声音。以这种方式思考语音被称为语音产生的源过滤理论。

The vocal tract is like a clarinet. The vocal folds are the reed, the source of the vibration. The column of air in the mouth is the column of air in the instrument, the filter of the vibration. And the speaker, of course, is the musician, changing the shape of the column of air to modulate the sound produced. Thinking about speech sounds in this way is called the source-filter theory of speech production.

它的工作原理如下。当空气从气管中流出并经过声带时，如果说话者将声带保持在正确的位置，声带就会开始振动。它们以每秒约 100 次（对于体型较大的成年男性）和每秒 300 次（对于儿童）的频率拍打和闭合。但除了这种基本的拍打动作之外，移动的声带中还有许多不同的子波纹。想象一下晾衣绳上被微风吹动的床单；它们不是僵硬地上下拍打，而是以复杂的图案波动。这些小波纹中的每一个都会为声音贡献自己的振动模式，从而产生“泛音”，或除了说话者声音的基本音调外，还有谐波。振动的基本频率，即基频，决定了音调，但泛音创造了不同声音的不同品质。

This is how it works. As air passes out of the trachea and over the vocal folds, if the speaker holds them in the correct position, the folds begin to vibrate. They flap open and closed at a frequency between about 100 times per second (for a large adult male) and 300 times per second (for a child). But on top of this basic flapping motion, there are many different subripples in the moving vocal folds. Think of sheets flapping in the breeze on a laundry line; they don’t flap stiffly up and down, but ripple in complicated patterns. Each of these little ripples contributes its own pattern of vibration to the sound, creating “overtones,” or harmonics, in addition to the basic pitch of the speaker’s voice. The basic rate of vibration, the fundamental frequency, determines the pitch, but the overtones create the different qualities of different sounds.

当声带振动时，它们会以同样复杂的方式引发声道内的空气振动。声道内的空气会过滤声带产生的声音的谐波结构。某些波纹（即某些谐波）会被放大，而其他波纹（其他谐波）则会被衰减。哪些谐波会被放大取决于声道内空气柱的形状。回想一下，不同的物体具有不同的特征振动模式，这取决于它们的大小和形状（例如手摇铃和教堂钟）。钟声的情况也适用于封闭的空气体。不同形状的空气体会以不同的频率振动。与特定声道形状的特征频率“一致”的谐波会被放大，而那些不一致的谐波会被减弱。说话者通过将舌头和嘴唇移动到不同的位置来控制过滤器，放大一些谐波并阻挡其他谐波。放大最强的频率称为共振峰。不同的元音有不同的共振峰结构。

As the vocal folds vibrate, they start the air inside the vocal tract vibrating in the same complex way. The air in the vocal tract then filters the harmonic structure of the sound produced at the vocal folds. Certain ripples (that is, certain harmonics) are amplified, and other ripples (other harmonics) are damped out. Which harmonics are amplified depends on the shape of column(s) of air inside the vocal tract. Recall that different objects have different characteristic patterns of vibration, depending on their size and shape (handbells vs. churchbells, for instance). What is true of bells is also true of enclosed bodies of air. Differently shaped bodies of air will tend to vibrate at different frequencies. Harmonics that are “in tune” with the characteristic frequencies of a particular vocal tract shape will be amplified, those that are not in tune will be reduced. The speaker controls the filter by moving the tongue and lips to different positions, amplifying some harmonics and blocking out others. The most strongly amplified frequencies are called formants. Different vowel sounds have different formant structures.

因此，根据舌头和嘴唇的形状，每个元音都有一种独特的、复杂的振动模式。振动从嘴唇传出，传播到世界各地。声波以每秒约 340 米的速度在空气中传播，直到它们撞击到专门用于接收它们的膜上，例如耳膜。

So, depending on the shape of the tongue and lips, each vowel sound has a characteristic, complex pattern of vibration. The vibration moves out past the lips, and propagates into the world. The sound waves travel through the air at the rate of about 340 meters per second, until they impinge on a membrane tuned to receive them, such as the eardrum.

听力

Hearing

耳朵分为三部分：外耳、中耳和内耳，如图1.7所示。外耳由可见的耳朵外壳（耳廓（英语：pinna ）和通向耳膜的耳道。耳廓有助于更好地捕捉声音，并在空间中定位声音。耳道保护耳膜，还有助于放大与语音相关的声音。声波沿着耳道中的空气传播，直到到达耳膜，耳膜开始振动。

The ear has three parts – the outer, middle, and inner ear, as shown in Figure 1.7. The outer ear consists of the visible shell of the ear (the pinna) and the ear canal leading down to the eardrum. The pinna helps to capture sounds better, and to locate sounds in space. The ear canal protects the eardrum, and also helps to amplify the sounds that are relevant for speech. Sound waves travel down the air in the ear canal until they reach the eardrum, which begins to vibrate.

中耳位于耳膜后面，是人体中最小的三块骨头（ 听小骨）。振动模式从耳膜通过听小骨传递到 中耳到内耳，耳蜗。通过中耳的这一额外传输阶段有助于放大非常小的声音，并降低非常大的声音。听力实际上是在耳蜗中发生的。耳蜗看起来像一个卷曲的蜗牛壳，比大理石还小。耳蜗分为上腔和下腔，内充满液体。至关重要的是，分隔两个腔的膜，即耳蜗膜，其长度并不均匀，而是逐渐变细。它一端厚 3 毫米，多骨，另一端薄（仅 1 毫米）且柔韧。整个膜上嵌入了微小的毛细胞，每根纤毛都附着在神经末梢上，在耳蜗液中轻轻摇动。

Behind the eardrum, in the middle ear, are the three smallest bones in the body (the ossicles). The patterns of vibration are transferred from the eardrum through the bones of the middle ear to the inner ear, the cochlea. This additional stage of transfer, through the middle ear, helps to amplify very soft sounds, and tone down very loud sounds. It is in the cochlea that hearing actually takes place. The cochlea looks like a curled-up snail shell, smaller than a marble. The cochlea is divided into upper and lower chambers, which are filled with fluid. Crucially, the membrane separating the two chambers, the cochlear membrane, is not uniform along its length, but is tapered. It is 3 millimeters thick and bony at one end and thin (just 1 millimeter) and flexible at the other. Embedded all along the membrane are tiny hair cells, the cilia, each attached to a nerve ending, and waving gently in the cochlear fluid.

当鼓膜振动时，中耳的骨头也会振动，进而引起内耳膜（耳蜗液体中的振动模式反映了说话者声道中产生的复杂声音模式——振动声带的基频以及不同元音的谐波特征 。

As the eardrum vibrates, the bones of the middle ear vibrate, which in turn cause the membrane of the inner ear (the oval window) to vibrate, starting up waves of vibration in the fluid of the inner ear. The patterns of vibration in the cochlear fluid mirror the complex pattern of sound created in the vocal tract of the speaker – the fundamental frequency of the vibrating vocal folds as well as the harmonic characteristics of the different vowels.

再次回想一下，不同大小和形状的物体倾向于以不同的频率振动。由于耳蜗膜的形状沿其长度变化（一端厚，另一端薄），膜上的不同位置对不同频率的振动有反应。厚端与低音调一致振动，薄端与高音调一致振动，而膜的中间部分与中音调一致振动。响应给定的振动模式，耳蜗膜上不同位置的纤毛被激活，向大脑发送有关传入波中存在的频率的信号。大脑将其感知到的频率信息重新组合成声音语言。

Again recall that objects of different size and shape tend to vibrate at different frequencies. Because the cochlear membrane varies in shape along its length (thick at one end, thin at the other) different places along the membrane respond to different frequencies of vibration. The thick end vibrates in tune to low-pitched sounds, the thin end in tune to higher-pitched sounds, and parts of the membrane in between vibrate in tune to mid-range sounds. In response to a given pattern of vibration, cilia at different places along the cochlear membrane are activated, sending signals to the brain about the frequencies present in the incoming wave. The brain recombines the frequency information it perceives into the sounds of language.

测量言语

Measuring speech

直到最近，语音学家的耳朵还是唯一可用于分析声音的工具，而且它们可能仍然是最精确和最可靠的。然而，从 20 世纪开始，语音学家开始发现让声音 可见。早期的设备，如示波器和声谱仪，使用麦克风将空气中的振动模式转换为电流变化模式。然后可以将这些变化显示在纸上或屏幕上，或通过电容器和电阻器组进行进一步测量和研究。

Until quite recently, the phonetician’s ears were the only instruments available for the analysis of sound, and they are probably still the most finely tuned and reliable. Beginning in the 1900s, however, phoneticians began to discover ways to make sound visible. Early devices, such as oscilloscopes and sound spectrographs, used a microphone to transfer patterns of vibration in the air into patterns of variation in electrical current. These variations could then be displayed – on paper or a screen – or passed through banks of capacitors and resistors for further measurement and study.

在二十一世纪初期，语音分析是通过计算机完成的。麦克风仍然将膜片的振动转换成电流的变化。但计算机将连续变化的声波转换成一系列数字；这称为模拟到数字 (A-to-D) 转换。（反向工作，让计算机、数字视频磁盘或数字音频播放器发出声音，是 D-to-A 转换。）

In the first part of the twenty-first century, speech analysis is done by computer. Microphones still convert the vibration of the membrane into variations in electrical current. But then computers convert a continuously varying sound wave into a series of numbers; this is called an analog-to-digital (A-to-D) conversion. (Working in the opposite direction, getting a computer, digital video disk, or digital audio player to make sound, is D-to-A conversion.)

一旦以数字格式表示和存储，声音文件就可以进行数学分析，以分离出不同的频率，并将分析结果以各种格式显示在屏幕上。图 1.8显示了三种常见且有用的显示。这三种显示都显示了“音素？”这句话（这些图形是在笔记本电脑上使用从互联网上免费下载的软件创建的。）

Once represented and stored in a digital format, sound files can be mathematically analyzed to separate out the different frequencies, and the results of the analysis displayed onscreen in various formats. Three common and useful displays are shown in Figure 1.8. All three show the utterance “A phoneme?” (These figures were created on a laptop computer, using software that was downloaded free from the internet.)

图 1.8A显示了波形；随时间变化的振动幅度（对应于响度）（在 y 轴上）被绘制出来（在 x 轴上）。话语的持续时间为一秒。波形上方是与波形的每个部分（大致）相对应的片段。请注意，元音 [o] 和 [i] 的振幅最大，阻塞音 [f] 的振幅最小，鼻音和非重音 [ ə ] 的振幅居中。图 1.8A还显示了波形各部分的特写视图（放大 12 倍）。在此放大倍数下，您可以看到摩擦音的随机噪声，以及每个元音特有的复杂、重复的振动模式。[o] 和 [i] 的模式不同，因此元音的质量也不同。波形是一种有用的表示方法，可用于查看话语的整体结构，识别不同类型的声音（停止，元音，我们可以使用语音识别技术（例如，鼻辅音）来检测语音信号中各个方面的持续时间。

Figure 1.8A shows a waveform; the varying amplitude (corresponding to loudness) of the vibrations (on the y-axis) is plotted over time (on the x-axis). The duration of the utterance is one second. Above the waveform are the segments (roughly) corresponding to each part of the waveform. Note that the vowels [o] and [i] have the greatest amplitude, the obstruent [f] the least, and the nasals and unstressed [ə] have an intermediate amplitude. Figure 1.8A also shows a close-up view (magnified 12 times) of portions of the waveform. At this magnification, you can see the random noisiness of the fricative, and the complex, repeating pattern of vibration that is characteristic of each vowel. The patterns for [o] and [i] are different, and thus the vowel qualities are different. The waveform is a useful representation for seeing the overall structure of an utterance, identifying the different types of sounds (stop, vowel, nasal consonant), and measuring the durations of different aspects of the speech signal.

由于音调变化承载着各种交流信息，语言学家经常希望测量说话过程中的基频。图 1.8B显示了这种计算的结果，音调轨迹。在此表示中，y 轴表示频率，x 轴表示时间。请注意，在 [f] 期间没有计算频率。由于此片段中没有声带振动，因此没有频率可测量。此音调轨迹显示此说话者的声音的基线基频约为 200 Hz，但此频率在话语结束时上升，呈典型的问题轮廓。

Because of all the communicative information carried by pitch changes, linguists often want to measure fundamental frequency over the course of an utterance. Figure 1.8B shows the result of such a computation, a pitch track. In this representation, the y-axis shows frequency, and time is on the x-axis. Notice that no frequency is calculated during the [f]. Since there is no vocal fold vibration during this segment, there is no frequency to measure. This pitch track shows that this speaker’s voice has a baseline fundamental frequency of about 200 Hz, but that this frequency rises toward the end of the utterance, in a contour typical of a question.

然而，这两个图都没有告诉我们太多有关元音音质的信息。我们可以从波形中看出 [o] 和 [i] 中的元音音质不同，因为模式看起来不同，但复杂的曲线并没有告诉我们太多有关组成泛音的信息，因此也没有告诉我们构成它们的声道形状。计算机可以进一步分析声波以分离其组成频率。这种分析的结果是频谱图，如图1.8C所示。

Neither of these figures, however, has told us much about the quality of the vowel sounds. We can see from the waveforms that the vowel qualities in [o] and [i] are different, because the patterns look different, but the complex squiggles don’t tell us much about the component overtones and thus about the vocal tract shape that made them. The computer can further analyze the sound wave to tease apart its component frequencies. The result of this analysis is a spectrogram, shown in Figure 1.8C.

与音高轨迹一样，x 轴表示时间，y 轴表示频率。但我们看到的不是单条线图，而是每个声音中存在的许多频率的复杂模式。我们希望分析结果告诉我们哪些频率最响亮，因为一旦我们知道哪些频率被放大了，我们可以使用已知公式来计算出产生这些频率的声道的（近似）形状。这个至关重要的第三维度，即振幅，由线条的暗度表示。某个频率的暗条表示该频率在声音中表现强烈。每个元音都有两个或三个最突出的频率模式，称为共振峰，高于说话者声带的基频。例如，在 [o] 元音中，我们看到共振峰位于约 1200 Hz。[i] 元音具有不同的共振峰结构，第二个共振峰位于约3000赫兹。

As with a pitch track, the x-axis represents time, and the y-axis shows frequency. But instead of a single line graph, we see a complicated pattern of the many frequencies present in each sound. We want our analysis to tell us which frequencies are the loudest, because once we know which frequencies have been amplified, we can use known formulas to figure out the (approximate) shape of the vocal tract that made them. This crucial third dimension, amplitude, is represented by the darkness of the lines. A dark bar at a certain frequency means that that frequency is strongly represented in the sound. Each vowel has a pattern of two or three most prominent frequencies, called formants, above the fundamental frequency of the speaker’s vocal folds. For example, during the [o] vowel, we see a formant at about 1200 Hz. The [i] vowel has a different formant structure, with the second formant at about 3000 Hz.

由于声道形状和共振峰结构之间的基本关系是已知的，我们可以从共振峰结构推断出我们看不到的舌头的位置。频谱图。对于不同的说话者，语音之间的一般关系保持不变：例如，[i] 中的第二共振峰始终高于 [o]。但由于每个人的声道大小和形状都是独一无二的，因此每个人的共振峰结构也是独一无二的我们也能够识别熟悉的声音，无论他们在说什么，在专家的手中，声谱图就像一个人的指纹。

Because the basic relationships between vocal tract shape and formant structure are known, we can infer the position of the tongue, which we can’t see, from the formant structure on the spectrogram. Across speakers, the general relationships between speech sounds remain the same: for example, the second formant is always higher in [i] than in [o]. But because every person’s vocal tract size and shape is unique, every person’s formant structure is unique too. We recognize familiar voices, regardless of what they’re saying, and in the hands of an expert, a spectrographic voice print is as unique as a fingerprint.

音系学

Phonology

关于语音，还有什么可说的呢？相当多。语言不仅存在于口中或耳朵中，还存在于大脑中。当我们从分析语音的物理方面转向研究其认知组织时，我们就从语音学转向了音系学。音系学永远无法完全脱离语音学，因为声音模式永远无法完全脱离其产生和听到的方式，而产生和感知总是受到总体语言组织的影响。

What else could there be to say about the sounds of speech? Quite a lot. Language isn’t just in the mouth or ears, but also in the brain. When we turn from analyzing the physical aspects of speech sounds to studying their cognitive organization, we move from phonetics to phonology. Phonology can never be completely divorced from phonetics, since sound patterns can never be completely separated from how they are produced and heard, and production and perception are always influenced by the overarching linguistic organization.

所有人类的声道和耳朵结构基本相同。那么为什么语言会有如此大的不同呢？某种程度上，这是因为它们使用了人类声道可能发出的不同声音。阿拉伯语使用咽舌擦音和小舌擦音，而英语则不使用。法语选择前圆元音（例如 [y] 和 [ œ ]），英语选择松高元音（[ ɪ ] 和 [ ʊ ]），西班牙语坚持使用基本发音 [i, e, a, o, u]。因此，在某种程度上，学习说一门新语言就是学习发出新的声音。

All human beings have basically the same structures in their vocal tracts and in their ears. So why are languages so different? To some extent, it is because they use different sounds from the repertoire of possible human vocal tract noises. Arabic uses pharyngeal and uvular fricatives, while English does not. French selects front round vowels (such as [y] and [œ]), English selects lax high vowels ([ɪ] and [ʊ]), and Spanish sticks to basic [i, e, a, o, u]. Thus, to some extent, learning to speak a new language is about learning to make new sounds.

但事情远不止如此。语言之间的差异不仅在于所使用的声音，还在于如何将这些声音组织成模式。例如，考虑一下英语和西班牙语的浊音阻塞音。

But there’s more to it than that. Languages differ not only in the sounds they use, but in how they organize those sounds into patterns. Consider, for example, the voiced obstruents of English and Spanish.

音素和音位变体

Phonemes and allophones

英语和西班牙语都有 [d] 和 [ð] 这两个音。例如，英语有den [d ɛ n] 和[ð ɛ n]，而西班牙语有 [dama] '女士' 和 [laðo] '侧面'。这两个音在两种语言中都有出现，但它们的分布（即声音出现的位置）不同，并且它们之间的差异所传达的信息也不同。

Both English and Spanish have the sounds [d] and [ð]. For example, English has den [dɛn] and then [ðɛn], while Spanish has [dama] ‘lady’ and [laðo] ‘side.’ Both sounds occur in both languages, but they differ in their distributions – that is, where the sounds appear – and in the information the difference between them conveys.

在一般美式英语中，有几对单词的区别仅在于一个单词带有 [d]，而另一个单词带有 [ð]：den [d ɛ n] 和[ ð ɛ n]、eider [a ɪ d ə r] 和either [a ɪ ð ə r]、bade [bed] 和bathe [beð]。在单词中同一位置上只有一个声音不同的单词对被称为最小对。（如果你把其中一个发音为 [ið ə r]，那么你就得到了一个接近最小对——两个几乎完全相同的单词。）最小对的存在意味着两个声音之间的差异是对比性的；把一个声音变成另一个声音，你就创造了一个意义的对比（也就是说，你造了一个不同的词）。如果两个声音是对比性的，它们的分布是不可预测的。如果我告诉你我正在想一个与when押韵的英语单词，并且以 [d] 或 [ð] 开头，您无法预测我想到的是哪个声音（或单词）。回顾表 1.1和1.3，它们给出了英语中不同声音的许多最小和接近最小对的例子。

In General American English, there are several pairs of words that differ only in that one has [d] where the other has [ð]: den [dɛn] and then [ðɛn], eider [aɪdər] and either [aɪðər], bade [bed] and bathe [beð]. Pairs of words that differ in only a single sound in the same position within the word are called minimal pairs. (If you pronounce either as [iðər], then you have a near-minimal pair – two words that are almost exactly alike.) The existence of minimal pairs means that the difference between the two sounds is contrastive; change one sound into another and you’ve created a contrast in meaning (that is, you’ve made a different word). If two sounds are contrastive, their distribution is unpredictable. If I tell you I’m thinking of an English word that rhymes with when, and starts with either [d] or [ð], you cannot predict which sound (or word) I have in mind. Look back at Tables 1.1 and 1.3, which give many examples of minimal and near-minimal pairs for different sounds in English.

现在将其与西班牙语的情况进行比较：

Now compare this to the situation in Spanish:

请注意，[d] 和 [ð] 在西班牙语中的分布不同。没有最小对——没有 [ðama] 与 [dama] 对立，没有 [lado] 与 [laðo] 对立。[d] 和 [ð] 之间的差异不是对比性的；[d] 与 [ð] 的差异在西班牙语中从未用于表示含义上的差异。相反，这两个声音具有不同的分布：只有 [d] 出现在起始位置，只有 [ð] 出现在元音后面。分布是可预测的。如果你知道上下文（单词中的位置或周围的声音），你可以预测 [d] 或 [ð] 会是用过的。

Notice that [d] and [ð] have a different distribution in Spanish. There are no minimal pairs – no [ðama] contrasting with [dama], no [lado] contrasting with [laðo]. The difference between [d] and [ð] is not contrastive; the [d] versus [ð] difference is never used in Spanish to signal a difference in meaning. Instead, the two sounds have different distributions: only [d] is found in initial position, and only [ð] is found following a vowel. The distribution is predictable. If you know the context (the position in the word, or the surrounding sounds), you can predict whether [d] or [ð] will be used.

当根据上下文可以预测两个不同的声音的发生时，我们说这两个声音是互补分布。（不是“好”意义上的“互补”，而是数学意义上的“互补”，即圆的一半是另一半的补音。）在西班牙语中，[d] 和 [ð] 的声音是互补分布的。（当考虑词首和词后以外的上下文时，情况会稍微复杂一些，但互补分布的原理是成立的。例如，[n] 后面只有 [d]，[r] 后面只有 [ð]。）

When the occurrence of two different sounds is predictable based on the context, we say that the two sounds are in complementary distribution. (Not “complimentary” in the sense of “nice,” but “complementary” in the mathematical sense that one half of a circle is the complement of the other half.) The sounds [d] and [ð] are in complementary distribution in Spanish. (The situation is slightly more complicated when contexts other than word-initial and postvocalic are considered, but the principle of complementary distribution holds. There is only [d] after [n], for instance, and only [ð] after [r].)

当一种语言中的两个声音形成最小对（即，如果它们的分布不可预测且具有对比性）时，这两个声音代表不同的音素。当一种语言中的两个声音呈互补分布（即，它们的分布可预测且不具有对比性）时，这两个声音是同一音素的变体。在英语中，[d] 和 [ð] 代表不同的音素。在西班牙语中，[d] 和 [ð] 是同一音素的变体。

When two sounds in a language form minimal pairs (that is, if their distribution is unpredictable and contrastive), those two sounds represent different phonemes. When two sounds in a language are in complementary distribution (that is, their distribution is predictable and noncontrastive), the two sounds are allophones of the same phoneme. In English, [d] and [ð] represent different phonemes. In Spanish [d] and [ð] are allophones of the same phoneme.

另一种说法是，音素是说话者认为是“相同”声音的一组声音的标签，而音位变体是根据产生该声音的上下文而发出的不同方式。对于西班牙语使用者来说，如果他/她稍加注意的话， [ð] 就是“一种说d的方式”。该语言的使用者知道什么时候需要 [d] 音位变体，什么时候适合使用 [ð] 音位变体。如果您正在学习西班牙语作为第二语言，您可能已经学过这种分布规则，例如“在元音之间，音 [d] 发音为 [ð]”。英语中没有将 [d] 和 [ð] 联系起来的规则。它们是独立的音素。

Another way to say this is that a phoneme is a label for a group of sounds that are perceived by the speaker to be the “same” sound, and the allophones are the different ways of pronouncing that sound depending upon the context in which it is produced. To the Spanish speaker, if he/she pays any conscious attention at all, [ð] is just “a way of saying d.” A speaker of the language knows when the [d] allophone is called for and when the [ð] allophone is appropriate. If you’re learning Spanish as a second language, you may have been taught this distribution as a rule, something like “the sound [d] is pronounced as [ð] between vowels.” There is no such rule relating [d] and [ð] in English. They are separate phonemes.

我们可以将这种情况描绘如下：

We may diagram the situation as follows:

音素用斜线表示，而同位音用括号表示。在同位音层面，英语和西班牙语的声音相同。在音素层面，英语有对比，而西班牙语没有对比。

Phonemes are indicated by slashes, while allophones are indicated by brackets. At the allophonic level, English and Spanish have the same sounds. At the phonemic level, English has a contrast where Spanish has none.

音位和同位音分布的差异给正在学习说另一种语言的人们带来了重大问题。以西班牙语为母语的人在学习英语时会很难区分den和then。对他/她来说，[d] 和 [ð] 算作同一个声音，所以他/她会倾向于这样听，并根据自己语言的原则来发音。西班牙语使用者可能倾向于将then说成 [d ɛ n] （使用词首同位音），将adder说成 [æðe.] （使用元音间的同位音）。这些不是随机错误，而是将第一语言的音系组织强加到第二语言单词上的结果 第二。

Differences in phonemic and allophonic distribution pose significant problems for speakers of one language who are learning to speak another. A native speaker of Spanish learning English will have trouble with the distinction between den and then. To him/her, [d] and [ð] count as the same sound, so he/she will tend to hear them that way, and to pronounce them according to the principles of his/her own language. A Spanish speaker may tend to say [dɛn] for then (using the word-initial allophone), and [æðe.] for adder (using the intervocalic allophone). These are not random errors, but a result of imposing the phonological organization of the first language onto the words of the second.

发现音素和音位变体

Discovering phonemes and allophones

本节将通过一个示例来说明如何确定声音属于一个或多个音素。本章末尾的练习中提供了更多练习示例。

In this section, we work through an example illustrating how to determine whether sounds belong to one or more phonemes. More practice examples can be found in the exercises at the end of the chapter.

分析必须从一些数据开始。要从头开始，语言学家必须找到所研究语言的使用者或使用者群体，询问单词和单词含义，然后转录它们。为了节省时间，我们将跳过第一步，然后假设数据已经准确收集。此外，在教科书练习中，由于时间和空间非常有限，读者必须相信作者选择的示例词具有代表性。在实际分析中，需要更大的数据集。

An analysis must begin with some data. To start from scratch, a linguist would have to find a speaker or group of speakers of the language under investigation, ask for words and word meanings, and transcribe them. In the interests of time, we’ll skip step one and assume that data have already been accurately collected. Further, in textbook exercises like this one, where available time and space is severely limited, the reader has to trust that the author has chosen the example words to be representative. In a real analysis, larger data sets would be required.

---

---

收集数据后，音位分析的第二步是确定一组可能被视为“相同”的声音。之所以选择这些声音，可能是因为它们相似，因为它们在许多其他语言中彼此相关，或者因为语言学家注意到他们怀疑这些声音的分布有限，是互补的。在这里，我们研究了斯瓦希里语中 [ ɔ ] 和 [o] 的分布（数据来自 Gleason 1955）。在表 1.5中，包含两个元音的斯瓦希里语单词分为两列，左侧是 [ ɔ ] 的单词，右侧是 [o] 的单词。有些单词同时包含这两个元音。

After the data have been collected, the second step in a phonemic analysis is to identify a set of sounds that are suspected to count as “the same.” They might be chosen because they are similar, because they have turned out to be related to each other in a number of other languages, or because the linguist notices a limited distribution that he/she suspects is complementary. Here, we examine the distribution of [ɔ] and [o] in Swahili (the data are from Gleason 1955). In Table 1.5, Swahili words containing the two vowels are listed in two columns, words with [ɔ] on the left and words with [o] on the right. Some words contain both.

收集并整理好数据集后，我们开始寻找模式。首先，寻找最小对。如果能找到，分析就完成了：这两个声音是对比性的，因此一定属于不同的音素。在这里，我们没有找到最小对：没有 [boma] 来匹配现有的单词 [b ɔ ma]；没有 [ ʃɔ ka] 来匹配单词 [ ʃ oka]。所以我们怀疑 [o] 和 [ ɔ ]的分布可能可以通过上下文预测。那么，下一步就是仔细观察每个声音出现的环境。列出每种情况下紧接在前面和后面的声音会有所帮助：这有助于将注意力集中在单词中最有可能相关的部分。查看紧接在前面的声音并不总是有效——有时声音会在一定距离外相互影响——但这是一个很好的起点。这在表 1.6中完成，其中空白表示每个声音的出现。符号# 表示单词的边缘，因此#__ 表示词首音，__# 表示词尾音。

Having collected and organized our data set, we proceed to search for patterns. First, look for minimal pairs. If any can be found, the analysis is finished: the two sounds are contrastive and therefore must belong to different phonemes. Here, we find no minimal pairs: there is no [boma] to match the existing word [bɔma]; no [ʃɔka] to match the word [ʃoka]. So we suspect that that the distribution of [o] and [ɔ] may be predictable by context. The next step, then, is to look closely at the environments in which each sound occurs. It can be helpful to list out the immediately preceding and following sounds in each case: this helps to focus on the part of the word most likely to be relevant. Looking at immediately adjacent sounds doesn’t always work – sometimes sounds can affect each other from some distance away – but it’s a good place to start. This is done in Table 1.6, where a blank indicates the occurrence of each sound. The symbol # indicates the edge of a word, so that #__ indicates a word-initial sound, and __# indicates a word-final sound.

查看表 1.6中的列表。尝试确定一个列表中的所有环境是否有共同点，但另一个列表中没有出现。有时模式会突然出现（就像这里可能出现的那样）；有时需要多加思考和努力。前面的声音似乎没有任何系统性的东西可以决定这个数据集中不同元音的选择。这两个列表都包含各种前面的鼻音和浊音和清音。特别是，[o] 和 [ ɔ ] 都可以跟在 [g] 后面，并且都可以出现在单词的开头。所以我们不能根据前面的声音预测会出现哪个元音。

Look at the lists in Table 1.6. Try to determine whether there’s anything that all the environments in one list have in common, but that doesn’t occur in the other list. Sometimes the pattern jumps out at you (as it probably does here); sometimes a little more thought and work is needed. There doesn’t seem to be anything systematic about the preceding sound that could condition the choice of a different vowel in this data set. Both lists contain a variety of preceding nasals and voiced and voiceless consonants. Particularly, both [o] and [ɔ] can follow [g], and both can occur at the beginning of a word. So we can’t predict which vowel will occur based on preceding sound.

另一方面，元音后面的声音有规律。元音 [ ɔ ] 只出现在 [m、n、ɲ或ŋ ] 之前，而元音 [o] 从不出现在这些辅音之前。这确定这两个声音处于互补分布状态。由于这两个声音处于互补分布状态，我们得出结论，它们属于同一个音素。

On the other hand, there is a pattern to the sounds that follow the vowel. The vowel [ɔ] only occurs before [m, n, ɲ, or ŋ], and the vowel [o] never occurs before any of these consonants. This establishes that the two sounds are in complementary distribution. Because the two sounds are in complementary distribution, we conclude that they belong to the same phoneme.

为了完成分析，我们需要说明哪些音位变体出现在何处的规则。说 [ ɔ ] 出现在 [m, n, ɲ , ŋ ] 之前，而 [o] 出现在 [t, d, k, g] 之前且位于词尾位置，是正确的，但仅仅给出一个声音列表会遗漏一些东西。音系分布很少（如果有的话）依赖于没有共同点的随机声音列表。音系分布依赖于自然类：具有共同点的声音集。声音 [m, n, ɲ , ŋ ] 具有共同点：它们都是鼻辅音。补集 [t, d, k, g, #] 没有任何共同属性：我们可以将其称为“其他地方的情况”。因此，我们可以如下表述我们的分析：[o] 和 [ ɔ ]在斯瓦希里语中处于互补分布，并且是一个音素的同位异音：[ ɔ ] 出现在鼻辅音之前，[o] 出现在其他地方。如图 1.9所示：

To complete the analysis, we need to state the rule for which allophones occur where. It would be correct to say that [ɔ] occurs before [m, n, ɲ, ŋ], and [o] occurs before [t, d, k, g] and in word-final position, but to just give a list of sounds is missing something. Phonological distributions seldom, if ever, depend on random lists of sounds with nothing in common. Phonological distributions depend on natural classes: sets of sounds that have something in common. The sounds [m, n, ɲ, ŋ] have something in common: they are all nasal consonants. The complement set [t, d, k, g, #] doesn’t have any property in common: we can call this the “elsewhere case.” Thus, we can state our analysis as follows: [o] and [ɔ] are in complementary distribution and allophones of one phoneme in Swahili: [ɔ] occurs before nasal consonants, [o] occurs elsewhere. This is diagrammed in Figure 1.9:

这个模式可以重新表述为一个概括：音素 /o/在鼻音前 发音为 [ ɔ ]。

The pattern can be restated as a generalization: the phoneme /o/ is pronounced [ɔ] before nasals.

请注意，我选择将音素标记为 /o/ 而不是 / ɔ /。在可能的范围内，语言学家会选择限制较少的变体（其他地方的情况）作为“基本形式”，而限制较多的变体作为“派生形式”。（出现在更多环境中并不一定意味着出现在更多单词中。[o] 和 [ ɔ ] 出现在斯瓦希里语单词列表中的词汇数量相等。）当选择其他地方的情况作为基本形式时，陈述概括会更简单。语句 a比语句b更简单：

Note that I’ve chosen to label the phoneme as /o/ rather than /ɔ/. To the extent that it’s possible, a linguist will choose the less restricted variant (the elsewhere case) as the “basic form,” and the more restricted variant as the “derived form.” (Occurring in more environments doesn’t necessarily mean occurring in more words. [o] and [ɔ] occur in an equal number of lexical items in this list of Swahili words.) It’s simpler to state a generalization when the elsewhere case is chosen as the basic form. Statement a is simpler than statement b:

并不总是能够明确地选择一个环境作为其他情况，但在所有条件相同的情况下，更简单的分析总是首选。人们应该始终尝试使分析尽可能简单和通用，并与数据保持一致。

It’s not always possible to definitively choose one environment as the elsewhere case, but all things being equal, the simpler analysis is always to be preferred. One should always try to make the analysis as simple and general as possible, consistent with the data.

---

---

音位学

Phonotactics

语音学家在描述一种未知语言时，首先要弄清楚的是该语言的声音清单：该语言使用什么声音？但他/她想弄清楚的第二件事是该语言使用哪些声音差异来编码单词之间的差异：音素是什么？接下来，语音学家试图通过识别可预测的互补分布模式来找出每个音素的不同音位变体。回答有关语言中声音的对比和可预测性的问题是语音学的主要工作。

The first thing a phonologist working to describe a previously unknown language wants to figure out is its inventory of sounds: what sounds does the language use? But the second thing he/she wants to figure out is which sound differences the language uses to encode differences between words: what are the phonemes? Next, a phonologist tries to figure out the different allophones of each phoneme by identifying predictable patterns of complementary distribution. Answering questions about contrast and predictability of sounds in a language is the main work of phonology.

语言不允许声音随机排列；相反，语言允许的声音序列是其结构的系统的和可预测的一部分。语言有音位学限制——对允许相邻出现或在单词中的特定位置出现的声音类型有限制。例如，在英语中，没有单词以序列 [tl] 开头。有 train和plane这样的单词，但没有tlane。这个序列完全可以发音，并且在其他语言中也出现过（例如纳瓦霍语 [tlee]‘night’）。如果两个声音在不同的音节中，例如At.lan.tic ，则该序列可以出现在英语中。但英语有音位学限制：* tl（“ tl是非法的”）在音节开头——因此也出现在单词开头。（星号用于指示在特定语言中不可能出现的任何序列。）

Languages do not allow random sequences of sounds; rather, the sound sequences a language allows are a systematic and predictable part of its structure. Languages have phonotactic constraints – restrictions on the types of sounds that are allowed to occur next to each other or in particular positions in the word. In English, for example, no word begins with the sequence [tl]. There are words like train and plane, but no tlane. This sequence is perfectly pronounceable and occurs in other languages (e.g. Navajo [tlee] ‘night’). The sequence can occur in English if the two sounds are in different syllables, as in At.lan.tic. But English has a phonotactic constraint: *tl (“t-l is illicit”) at the beginning of a syllable – and therefore at the beginning of a word, as well. (The asterisk is used to indicate any sequence that is not possible in a given language.)

英语的另一个音位限制限制了双元音 [a ʊ ]后面可以发出的声音。我们有许多单词，例如out、crowd、town、mouse、couch、south和rouse。但是没有单词^* aup、^* auk、^* aub、^* awm、^* aug等。如果你列出所有你能想到的 [a ʊ ] 单词，你会发现 [a ʊ ]后面可以发出的声音只能是用舌尖和舌叶发出的声音：[t]、[d]、[s]、[z]、[n]、[ θ ]、[ð]、[ ʧ ]、[ ʤ ]（如果将owl和hour作为一个音节发声，则可能还有 [l] 和 [r] ）。因此，表达这种英语音位约束的一种方式是“双元音 [a ʊ ] 后面只能跟着 [t]、[d]、[s]、[z]、[n]、[ θ ]、[ð]、[ ʧ ]、[ ʤ ]”。但我们可以做得更好。我们想要捕捉到这样的概括：这些声音不是随机的列表，而是都具有一些共同的属性。表达这种约束的更好方法是“双元音 [a ʊ ] 后面只能跟着用舌头前部发出的辅音。”音位约束很少（如果有的话）针对随机的声音集合。相反，它们几乎总是针对具有一个或多个共同语音属性的声音组。这样的组被称为自然类别。

Another phonotactic constraint of English limits the sounds that are allowed to follow the diphthong [aʊ]. We have many words like out, crowd, town, mouse, couch, south, and rouse. But there are no words ^*aup, ^*auk, ^*aub, ^*awm, ^*aug, etc. If you make a list of all the [aʊ] words you can think of, you will discover that the only sounds allowed to follow [aʊ] are those made with the tongue tip and blade: [t], [d], [s], [z], [n], [θ], [ð], [ʧ], [ʤ] (and possibly [l] and [r], if you say owl and hour as one syllable). So one way of stating this English phonotactic constraint would be “The diphthong [aʊ] can only be followed by [t], [d], [s], [z], [n], [θ], [ð], [ʧ], [ʤ].” But we can do better. We want to capture the generalization that these sounds are not a random list, but all share some property. A better way to state this constraint is “The diphthong [aʊ] can only be followed by a consonant made with the tongue front.” Phonological constraints seldom (if ever) target random collections of sounds. Rather, they almost always target groups of sounds that have one or more phonetic properties in common. Such groups are called natural classes.

跨语言最常见的音位限制之一是“鼻音必须与下列发音一致停止。”想想那些有序列的英语单词鼻音 + 口塞音：camp，hamper，bombard，paint，intelligent，wind，window。虽然也有一些例外，但一般的模式是双唇塞音 [p] 和 [b] 前面是双唇鼻音 [m]，而齿龈塞音 [t] 和 [d] 前面是齿龈鼻音 [n]。那么带有 [k] 和 [g] 的单词呢？我们将这些单词拼写为nk和ng ，但如果你仔细听自己对think和linguistics等单词的发音，你会发现这些序列实际上是 [ ŋ k] 和 [ ŋ g] （软腭鼻音 [ ŋ ] 位于软腭塞音 [k] 和 [g] 之前），因此也遵循该限制。 （许多拼写以ng结尾的单词，例如sing和rang，实际上仅以鼻音 [ ŋ ] 结尾，而根本没有 [g] 塞音。）

One of the most common phonotactic constraints across languages is “Nasals must agree in place of articulation with a following stop.” Think of English words that have a sequence of nasal + oral stop: camp, hamper, bombard, paint, intelligent, wind, window. Though there are some exceptions, the general pattern is for the bilabial stops [p] and [b] to be preceded by the bilabial nasal [m] and for the alveolar stops [t] and [d] to be preceded by the alveolar nasal [n]. What about words with [k] and [g]? We spell these words nk and ng, but if you listen carefully to your pronunciation of words like think and linguistics, you’ll realize that these sequences are really [ŋk] and [ŋg] (the velar nasal [ŋ] preceding the velar stops [k] and [g]), and thus also follow the constraint. (Many words whose spelling ends in ng, like sing and rang, actually end in just the nasal [ŋ], with no [g] stop at all.)

有许多不同的语音维度可以定义自然类别。浊音在音位约束中通常很重要。一个常见的约束是，阻塞音序列必须全部浊音或全部清音。在俄语中有 [vzbut ʃ ki] ‘责骂’ 和 [fspleska] ‘泼’ 等词，但没有^* [fzput ʃ gi]。英语中有 [s ɪ ks θ s]，但没有^* [s ɪ ks θ z]。在日语本土词中（从中文或其他语言借来的词除外），每个词只能有一个浊音阻塞音。因此，有 [saki]“酒”，[kaze]“风”，[zaseki]“座位”（每个词都有一个或多个清音阻塞音），但没有^* [gaze] 或^* [guda]。

There are many different phonetic dimensions that can define natural classes. Voicing is often important in phonotactic constraints. A common constraint is that sequences of obstruents must be either all voiced or all voiceless. In Russian, there are words like [vzbutʃki] ‘scolding’ and [fspleska] ‘splash,’ but not ^*[fzputʃgi]. In English, we have [sɪksθs], but not ^*[sɪksθz]. In native Japanese words (borrowings from Chinese or other languages are exempt), there can be only one voiced obstruent per word. So there are words like [saki] ‘wine,’ [kaze] ‘wind,’ and [zaseki] ‘seat’ (each with one or more voiceless obstruents), but there is no ^*[gaze] or ^*[guda].

音位限制通常与音节结构有关。许多语言都对它们允许的音节类型有所限制。所有语言都允许 CV（辅音-元音）类型的音节 - 即有声母和音节核心，但没有尾音。这是标记最少且最常见的音节类型。有些语言，如西非语言Senufo，只允许使用 CV 音节。对 CV 音节的偏好源于两个音位学限制：“音节必须有首音”（V 前有一个 C）和“音节不能有尾音”（V 后没有 C）。

Phonotactic constraints are often related to syllable structure. Many languages put constraints on the types of syllables they will allow. All languages allow syllables of the CV (Consonant–Vowel) type – that is, with an onset and a nucleus, but without a coda. This is the least marked and most common type of syllable. Some languages, like the West African language Senufo, allow only CV syllables. The preference for CV syllables results from two phonotactic constraints: “Syllables must have onsets” (a C before the V), and “Syllables must not have codas” (no C after the V).

有些语言只有其中一个限制。例如，夏威夷语不允许使用尾音，但允许使用没有声母的音节。因此，它允许使用 V 和 CV 类型的音节，但绝不允许使用 CVC 类型的音节。夏威夷语还以拥有最小的音节库而闻名 – 只有五个元音 [i、e、a、o、u] 和八个辅音 [p、k、?、l、w、n、m、h]。这样一共有 45 个可能的音节（5 个 V 和 40 个 CV 组合）来编码该语言中的数千个单词。如何使用 45 个音节来组成数千个单词？夏威夷语单词可以非常非常长，并且有很多重复！例如，该州的鱼被称为 [hu.mu.hu.mu.nu.ku.nu.ku.a.pu.a. ʔ a]。请注意，每个音节都是 CV 或 V，并且 [ ʔ ] 起常规辅音的作用。

Some languages have only one of these constraints. Hawai‘ian, for example, does not allow codas, but does allow syllables that don’t have onsets. Therefore, it allows syllables of the type V and CV, but never CVC. Hawai‘an is also famous for having one of the smallest segmental inventories around – only five vowels [i, e, a, o, u] and eight consonants [p, k, ?, l, w, n, m, h]. That gives a grand total of 45 possible syllables (5 V and 40 CV combinations) to encode the thousands of words in the language. How do you use 45 syllables to make thousands of words? Hawai‘ian words can be very, very long, with lots of repetition! The state fish, for example, is called the [hu.mu.hu.mu.nu.ku.nu.ku.a.pu.a.ʔa]. Note that every syllable is either CV or V, and that [ʔ] functions as a regular consonant.

其他语言允许使用 CV 和 CVC，但不允许使用 VC。这些语言中的每个音节都必须有首音。阿拉伯语和德语属于这种类型。书面上以元音开头的单词实际上是用初始喉塞音——例如，德语中的[ ʔ apf ɛ l] '苹果'，阿拉伯语中的 [ ʔ al] '这个'。

Other languages allow CV and CVC but not VC. Every syllable in these languages must have an onset. Arabic and German are languages of this type. Words that are spelled with initial vowels in writing are actually pronounced with an initial glottal stop – for example, [ʔapfɛl] ‘apple’ in German, and [ʔal] ‘the’ in Arabic.

许多允许音节有尾音的语言（CVC 和 VC）限制了可以出现在那里的声音。这些限制被称为尾音约束。在泰语中，[p]、[p ^h ] 和 [b] 可以出现在声母中，但只有 [p] 可以出现在尾音中。在日语中，唯一​​允许的尾音辅音是鼻音或双音节的前半部分：[nip.pon]“日本”，[ ʃ im.bun]“报纸”，[gak.ko]“学校”。声母有限制（例如英语的声母中没有 [ ŋ ]），但这种情况要少见得多。最后，许多语言最多允许尾音中有一个辅音，声母中也有一个辅音。英语很少允许使用 CCVCC 甚至 CCCCVCCC 词，例如shrink或sprints。俄语单词如 [fspleska]（连续四个辅音）则更为少见。

Many languages that allow syllables to have codas (CVC and VC) limit the sounds that can occur there. These restrictions are called coda constraints. In Thai, [p], [p^h], and [b] may occur in onsets, but only [p] may appear in the coda. In Japanese, the only coda consonants allowed are nasals or the first half of geminates: [nip.pon] ‘Japan,’ [ʃim.bun] ‘newspaper,’ [gak.ko] ‘school.’ Constraints on onsets exist (e.g. no [ŋ] in onsets in English), but are much rarer. Finally, many languages allow at most one consonant in the coda and one in the onset. English is rather rare in allowing CCVCC, or even CCCVCCC words, like shrink or sprints. Russian words like [fspleska] (four consonants in a row) are rarer still.

---

---

交替和异形体

Alternation and allomorphs

音系学是关于对比和可预测性的。我们想知道一种语言中哪些声音是对比音素，我们希望能够预测不同音素在什么语境中音素的变体会出现。有时我们通过研究单词列表来发现不同类型的声音分布。有时我们观察借词的情况。另一种方法是研究交替。

Phonology is about contrast and predictability. We want to know what sounds in a language are contrastive phonemes, and we want to be able to predict the contexts in which the different allophones of a phoneme appear. Sometimes we do this by studying lists of words in order to discover different kinds of sound distributions. Sometimes we look at what happens to borrowed words. Another way is by studying alternations.

交替是指同一词素根据上下文以不同方式发音。词素是你心理词典中的任何条目；它可以是单词或单词的一部分，如前缀（添加到单词的开头）或后缀（添加到单词的末尾）。例如，在英语中，cat是一个词素，复数后缀 - s也是一个词素。当我们将不同的词素组合在一起以创建更大的单词时，我们会创建不同的上下文。根据上下文，会发音不同形式的词素，即不同的词形变体。当词素的组合创建了违反音位学的新环境时，相同的音位学限制（不允许词素内出现某些声音序列）也会导致交替限制。

An alternation is seen when the same morpheme is pronounced different ways depending on the context. A morpheme is any entry in your mental dictionary; it can be a word or a part of a word, like a prefix (added to the beginning of a word) or a suffix (added to the end of a word). For example, in English, cat is a morpheme, and so is the plural suffix -s. As we put different morphemes together to create larger words, we create different contexts. Depending on the context, different forms of the morpheme, different allomorphs, are pronounced. The same phonotactic constraints that disallow certain sequences of sounds within morphemes also give rise to alternations when the combination of morphemes creates new environments that violate phonotactic constraints.

我们已经看到，在英语的词素中，鼻音与随后的塞音的发音位置一致：remember camp、wind和think。因为这些鼻音位于单个词素中，所以我们从未看到过它们的发音方式的变化。但是，有些以鼻音结尾的英语前缀在添加到以不同声音开头的单词中时发音会有所不同。考虑否定前缀in -。在元音或齿龈音前，in - 发音为 [ ɪ n]：inedible、inaudible、indelible、inolerable、insufferable。但在双唇音 [m]、[b] 或 [p] 之前，它变成 [ ɪ m]：immovable、inbalance、irreparial。在 [k] 或 [g] 之前，发音为 [ ɪŋ ]。我们不改变拼写，因为英语没有单独的字母表示软腭鼻音，但听听你对像invalid 这样的单词的发音，你会发现鼻音是软腭而不是齿龈。音位约束“鼻音必须与随后的塞音一致”不仅适用于词素（如camp和wind），也适用于词素合二为一。

We saw that, within morphemes in English, nasals agree with the place of articulation of a following stop: remember camp, wind, and think. Because these nasals are within a single morpheme, we never see alternations in how they are pronounced. But some English prefixes that end with a nasal are pronounced differently when the prefix is added to words beginning with different sounds. Consider the negative prefix in-. Before a vowel or an alveolar, in- is pronounced [ɪn]: inedible, inaudible, indelible, intolerable, insufferable. But before the bilabials [m], [b], or [p], it becomes [ɪm]: immovable, imbalance, impartial. And before [k] or [g], it’s pronounced [ɪŋ]. We don’t change the spelling, because English doesn’t have a separate letter for the velar nasal, but listen to your pronunciation of a word like incorrect and you’ll realize the nasal is velar rather than alveolar. The phonotactic constraint “Nasals must agree in place of articulation with a following stop” applies not just within morphemes (like camp and wind) but also when morphemes are combined.

自然类别帮助我们定义要改变的声音集、改变本身以及改变发生的环境。这在以下情况下可以清楚地看到西班牙语发音停顿和上面描述的擦音。我们在那里注意到 /d/ 在西班牙语中有两个同位音：[d] 在起始位置，[ð] 跟在元音后面。但这种变化并不仅限于 /d/。它还影响 /b/ 和 /g/。每当浊塞音 [b, d, g] 跟在元音后面时，它就会变为相应的浊擦音 [ β ,ð, ɣ。

Natural classes help us to define the set of sounds that is targeted for alternation, the change itself, and the environments where the change takes place. This can be clearly seen in the case of the Spanish voiced stops and fricatives described above. We noted there that /d/ has two allophones in Spanish: [d] in initial position, and [ð] following a vowel. But the change is not limited to /d/ alone. It also affects /b/ and /g/. Whenever a voiced stop [b, d, g] follows a vowel, it changes to the corresponding voiced fricative [β, ð, ɣ].

发生变化的声音集是一个自然类：浊音塞音。它们都经历了相同的变化；它们变成了浊擦音。变化发生的环境可以用一个语音参数来定义：在具有开放声道和连续气流的声音之后，如元音 [s] 和 [r]。

The set of sounds that undergoes change is a natural class: voiced stops. They all undergo the same change; they become voiced fricatives. And the environment in which the change occurs can be defined by a single phonetic parameter: after sounds that have an open vocal tract and continuous flow of air, like vowels, [s], and [r].

音系交替的类型

Types of phonological alternations

我们发现某些类型的音系过程在各种语言中反复出现。（请参阅第 8 章，从历时（跨时间）角度了解以下更多交替示例。）

There are certain types of phonological processes that we find over and over again across languages. (See Chapter 8 for more examples of the following alternations in a diachronic (across time) perspective.)

同化

最常见的交替类型是同化：两个不同的声音变得更相似。同化可能是局部的，即两个声音彼此相邻，也可能是远距离的，即两个声音即使通过其他中间片段似乎也会相互影响。

The most common type of alternation is assimilation: two sounds that are different become more alike. Assimilation may be either local, when the two sounds are next to each other, or long-distance, where two sounds seem to affect each other even through other intervening segments.

浊化同化是跨语言的一种非常常见的变化。在英语中，复数后缀与前一个塞音一致：两只猫[kæts] 和两只狗[dagz]。另一个例子是俄语。俄语中表示“来自”的单词是 [ot] 或 [od]：[od vzbút ʃ ki] “来自责骂”，[ot fspléska] “来自水花”。

Voicing assimilation is one very common alternation across languages. In English, the plural suffix agrees in voicing with a previous stop: two cats [kæts] and two dogs [dagz]. Another example is seen in Russian. The Russian word for ‘from’ is either [ot] or [od]: [od vzbútʃki] ‘from a scolding,’ [ot fspléska] ‘from a splash.’

局部同化也可能影响发音部位。英语 中的否定前缀- 就是鼻音同化发音部位的例子，该部位发音部位与随后的塞音相似（不雅与不可能）。另一个例子来自西非语言特维语。

Local assimilation may also affect place of articulation. The English negative prefix in- is an example of a nasal assimilating in place of articulation to a following stop (indecent vs. impossible). Another example comes from Twi, a language of West Africa.

这在特维语中表示‘不’的词素有三种不同的变体。它总是以鼻辅音的形式发出，但鼻音会取代它旁边的任何辅音的发音位置 – 双唇音旁边是双唇音，软腭音旁边是软腭音，等等。

The morpheme that means ‘not’ in Twi has three different allomorphs. It is always produced as a nasal consonant, but the nasal takes on the place of articulation of whatever consonant it’s next to – bilabial next to bilabial, velar next to velar, etc.

最极端的局部同化类型是完全同化：两个相邻的声音变得相同。许多西非语言都存在相邻元音的完全同化。在例如，约鲁巴语中 [ow o ]‘钱’加上 [ e po]‘石油’变成 [ow ee po]‘石油钱’。伊博语中，[nwok e ]“人”加上[ a ]“限定词”变成[nwo kaa ]“那个人”。英语前缀/ ɪn /在某些词中也可以完全同化。在大多数辅音前，鼻音只同化到发音位置，如inpartial。但在[l]或[r]之前，鼻音的同化是完全的：in + relevant变成irlevant，in + legible变成illegible。

The most extreme type of local assimilation is complete assimilation: two sounds that are next to each other become identical. Complete assimilation of adjacent vowels is found in many West African languages. In Yoruba, for example [owo] ‘money’ plus [epo] ‘oil’ becomes [oweepo] ‘oil money.’ In Igbo, [nwoke] ‘man’ plus [a] ‘determiner’ becomes [nwokaa] ‘that man.’ The English prefix /ɪn/ can also undergo complete assimilation in certain words. Before most consonants, the nasal assimilates just to the place of articulation, as in impartial. But before [l] or [r], the assimilation of the nasal is total: in + relevant becomes irrelevant, in + legible becomes illegible.

同化也可以发生在远距离。两个不直接相邻的音段可能会共享一些语音特性。元音和谐是长距离同化的典型例子。例如， 土耳其语中，单词中的元音必须全部在前或全部在后，高元音必须全部圆音或全部不圆音。因此，表示（大致）“of”的后缀具有不同的同位异形，具体取决于前一个音节中元音的质量：[ip-in]“绳索的”，[pul-un]“邮票的”。在和声系统中，即使辅音介入，元音也会相互同化。辅音和声也存在，尽管它们是更加稀有。

Assimilation can also take place long-distance. Two segments that are not immediately adjacent may come to share some phonetic property. Vowel harmony is the prime example of long-distance assimilation. For example, in Turkish, the vowels in a word must be all front or all back, and the high vowels must be all round or all unround. The suffix meaning (roughly) ‘of’ therefore has different allomorphs, depending on the quality of the vowel in the preceding syllable: [ip-in] ‘of the rope,’ [pul-un] ‘of the stamp.’ In harmony systems, the vowels assimilate to each other, even though consonants intervene. Consonant harmonies also exist, though they are rarer.

异化

同化的反义词是异化。两个相似的发音变得不同。异化的动机之一可能是发音的容易程度。两个相似但不完全相同的发音似乎特别难以在相邻的位置正确发音。（这一原则构成了许多绕口令的基础，例如第六位酋长的第六只羊。）解决这个问题的一个方法是使相似的发音相同；另一个解决方案是使它们更加不同。异化的一个例子可以在历史中找到希腊语。在古希腊语中，表示学校的单词是 [sxolio]，有两个相邻的清擦音。在现代希腊语中，它变成了 [skolio]，取而代之的是擦音-塞音序列。异化也可能帮助听者意识到存在两个音段，而不仅仅是一个。在塞茨瓦纳语（在博茨瓦纳使用）中，浊塞音在（浊）鼻音后变为清音，也许是为了使塞音的存在更加明显：[bona]“see”变成[mpona]“see我。'

The opposite of assimilation is dissimilation. Two sounds that are similar become different. One impetus for dissimilation may be ease of articulation. Two sounds that are similar but not exactly the same seem to be particularly difficult to pronounce correctly right next to each other. (This principle forms the basis of many tongue twisters, such as the sixth sheik’s sixth sheep.) One solution to the problem would be to make the similar sounds identical; another solution is to make them more different. An example of dissimilation is found in the history of Greek. In Ancient Greek, the word for school was [sxolio], with two adjacent voiceless fricatives. In Modern Greek, this has become [skolio], with a fricative–stop sequence instead. Dissimilation may also help hearers to realize that two segments are present, not just one. In Setswana (spoken in Botswana), voiced stops become voiceless after a (voiced) nasal consonant, perhaps to make the presence of the stop more obvious: [bona] ‘see’ becomes [mpona] ‘see me.’

插入

另一种常见的交替类型是插入（音系学家称之为插入音）。插入通常与音节结构有关，正如我们上面所见，插入元音来断开一连串的辅音。这种情况可能发生在外来词中（如日语和夏威夷语），也可能发生在词素组合在一起时。在英语中，当我们想在已经以 [s] 或 [z] 结尾的单词中添加复数后缀 [ z ] 时，我们会插入元音 [ ɨ ] 来断开两个高音擦音：一个 dress [dr ɛ s]，两个 dress [dr ɛ s ɨ z]。

Another common type of alternation is insertion (which phonologists call epenthesis). Insertion is usually related to syllable structure, as we saw above, when vowels are inserted to break up strings of consonants. This can happen in loan words (as in Japanese and Hawai‘ian) or when morphemes come together. In English, when we want to add the plural suffix [z] to a word that already ends in [s] or [z], we insert a vowel [ɨ] to break up the two high-pitched fricatives: one dress [drɛs], two dresses [drɛsɨz].

删除

当然，插入的反义词是删除。一种语言可能不是用元音打断辅音序列，而是选择删除其中一个辅音（如pneumonia中首字母 [p] 的消失）。Grand以 [nd] 结尾，mother以[m] 开头，但当两个单词放在一起时，中间的 [d] 通常会被删除：[grænm ʌ ð ə .]。（[n] 可能会被同化为 [m]，从而变成 [græmmm ʌ ð ə r]。）Lardil（一种澳大利亚语言）会删除三个或三个以上音节的单词的末尾元音：[yalulu] ‘flame’ 变成 [yalul]。

The opposite of insertion, of course, is deletion. Instead of breaking up a sequence of consonants with a vowel, a language may choose to delete one of the consonants (as in the loss of the initial [p] in pneumonia). Grand ends with [nd], mother starts with [m], but when the two words are put together, the medial [d] is usually deleted: [grænmʌðə.]. (The [n] may then assimilate to the [m], resulting in [græmmmʌðər].) Lardil (a language spoken in Australia) deletes final vowels from words of three or more syllables: [yalulu] ‘flame’ becomes [yalul].

缓和与强化

另一种变化是声音减弱，即声音变得更柔和或更弱。停止更改为擦音，擦音变为近似值。西班牙语交替 [b, d, g] 变成 [ β , ð, ɣ ] 是轻化音的一个例子。轻化音的反义词是强化音。在这里，摩擦音变成塞音，例如基库尤语（东非）“鼻后硬化”：[ ɣ ora] '购买，' [ ŋ goreet ɛ ] '我已经买了。'

Another type of change is lenition – sounds become softer or weaker. Stops change to fricatives, fricatives change to approximants. The Spanish alternation [b, d, g] becomes [β, ð, ɣ] is an example of lenition. The opposite of lenition is fortition. Here, fricatives change into stops, as in Kikuyu (East Africa) “postnasal hardening”: [ɣora] ‘buy,’ [ŋgoreetɛ] ‘I have bought.’

换位和重复

其他不太常见的变化包括换位和重复。换位意味着改变声音的顺序。例如，英语单词horse曾经是 [hros]，后来才变成 [ro] 序列。重复意味着复制。在英语中，我们有时会复制单词的一部分来传达贬义、小意思：teeny-tiny、itsy-bitsy，或者更轻蔑的syntax-schmintax。但在其他语言中，它是语音系统的常规组成部分。在Manam（南岛语），[gara] 的意思是“刮”，[gara-gara-si] 的意思是“刮”。您可能还记得夏威夷语 [humuhumunukunukuapua ʔ a] 和 Lardil [ ʧ umpu ʧ umpu]。

Other less common alternations include metathesis and reduplication. Metathesis means switching the order of sounds. For example, the English word horse used to be [hros], before the [ro] sequence was metathesized. Reduplication means copying. In English, we sometimes copy parts of words to convey a pejorative, diminutive sense: teeny-tiny, itsy-bitsy, or the more dismissive syntax-schmintax. But in other languages it’s a regular part of the sound system. In Manam (Austronesian), [gara] means ‘scrape’ and [gara-gara-si] means ‘scraping it.’ You may also recall Hawai‘ian [humuhumunukunukuapuaʔa] and Lardil [ʧumpuʧumpu].

发现替代

Discovering alternations

进行音位变换分析的第一步是确定需要考虑的词素的不同形式。英语中过去时后缀有三种形式：[t] 如reached [rit ʃ t]、[d] 如mowed [mo:d] 和 [ ɨ d] 如rotted [ra ɾɨ d]。表 1.7列出了带有每个后缀的各种单词。音标显示了未加后缀的动词的发音。（音标很宽泛，不包括与变换无关的非对比细节。）

The first step in doing an analysis of phonological alternation is to identify the different forms of the morpheme that need to be accounted for. The past-tense suffix in English has three forms: [t] as in reached [ritʃt], [d] as in mowed [mo:d], and [ɨd] as in rotted [raɾɨd]. Table 1.7 gives a list of various words that take each suffix. The transcriptions show the pronunciation of the unsuffixed verb. (Transcription is broad, and noncontrastive details not relevant to the alternation are not included.)

作为分析员，我们希望发现这种变化是否受音系环境的影响。因此，下一步（与同位异音分析一样）是列出每个同位异音出现的环境。事实上，进行同位异音分析的步骤如下：音素/变体分析和词素交替分析基本相同。区别在于关注点。在音素/变体分析中，语言学家关注的是整个词汇中的声音分布（当然，在教科书示例或练习中，人们只能关注代表性单词的子集）。在词素交替分析中，语言学家关注的是特定单词或词缀的变体形式。然而，在这两种情况下，总体目标是相同的：发现声音结构的哪些方面是对比的，哪些是可预测的根据环境。

As analysts, we want to discover if the alternation is conditioned by phonological environment. Therefore, the next step (as with an allophone analysis) is to list the environments where each allomorph occurs. In fact, the steps in doing a phoneme/allophone analysis and a morpheme alternation analysis are pretty much the same. The difference is one of focus. In a phoneme/allophone analysis, the linguist is looking at the distribution of sounds across the entire lexicon (though of course in a textbook example or exercise one can only look at a subset of representative words). In a morpheme alternation analysis, the linguist is looking at the variant forms of a specific word or affix. In both cases, however, the overall goal is the same: to discover which aspects of sound structure are contrastive and which are predictable according to the environment.

英语过去时后缀变体的环境如表 1.8所示。由于交替词素是后缀，我们可以假设词干的最后一个音决定了变化。

The environments for the variants of the English past-tense suffix are shown in Table 1.8. Since the alternating morpheme is a suffix, we can begin with the hypothesis that it is the final sound of the stem that conditions the change.

我们立即注意到，同位素变体呈互补分布：没有一个声音出现在多个列表中。我们还可以注意到，每个列表都形成一个可以用语音术语定义的类：[ ɨ d] 出现在齿龈塞音之后，[t] 出现在所有其他清音之后，[d] 出现在所有其他浊音之后聲音。

We can immediately note that the allomorphs are in complementary distribution: no sound occurs in more than one list. We can also note that each list forms a class that can be defined in phonetic terms: [ɨd] occurs following alveolar stops, [t] occurs following all other voiceless sounds, and [d] occurs following all other voiced sounds.

此外，我们可以看到音位约束在激发交替方面的作用。插入 [ ɨ ] 是为了拆分一串辅音（[td] 或 [dd]），否则这些辅音在词尾是无法发音的。清辅音后从 [d] 变为 [t] 使序列符合要求阻塞音群在浊音上一致的约束。如上所述，音位约束排除了标记结构：“固定”的音群要么难以听清（例如 [hændd] 中的第二个 [d]），要么难以发出（例如在 [lpd] 等辅音群中间两次切换浊音）。

Further, we can see the activity of phonotactic constraints in motivating the alternation. The [ɨ] is inserted to break up a cluster of consonants ([td] or [dd]) that would otherwise be unpronounceable at the end of a word. The change from [d] to [t] after voiceless consonants brings the sequence into line with the constraint that requires obstruent clusters to agree in voicing. As was noted above, the phonotactic constraints rule out marked structures: the clusters that are “fixed up” are either hard to hear (like a second [d] in [hændd], or hard to produce (like switching voicing twice in the middle of a consonant cluster like [lpd]).

为了完成分析，我们需要确定词素的基本形式，并陈述得出每种变体形式的概括。在这里，我们可以放心地选择 [d] 作为其他地方的情况，因为采用 [d] 的声音集合是最多样化的：元音、响音和浊音。（这个集合实际上构成了自然类浊音，但这或多或少是偶然的：后缀必须跟在某个东西后面，如果一个声音不是无声的，它就必须是有声的。）我们可以通过首先应用最具体情况的规则来解决确保 [t] 和 [d] 获得其自己的特殊形式的问题。（这种首先考虑最具体情况，从而将其排除在外，以便可以更普遍地做出后续陈述的想法最初是由古印度语法学家 P āṇ ini 提出的，现在仍被称为 P āṇ ini 定理。）

To finish the analysis, we need to decide on the basic form of the morpheme, and state the generalizations that derive each of the variant forms. Here, we can be secure in choosing [d] as the elsewhere case, as the set of sounds that take [d] is the most diverse: vowels, sonorants, and voiced obstruents. (This set does in fact form the natural class of voiced sounds, but that is more or less accidental: a suffix has to follow something, and if a sound isn’t voiceless it has to be voiced.) We can take care of the problem of making sure that [t] and [d] get their own special form by applying the rule for the most specific case first. (This idea of taking the most specific case first, thus getting it out of the way so that succeeding statements could be made more generally, is originally due to the ancient Indian grammarian Pāṇini, and is still known as Pāṇini’s Theorem.)

构成英语过去时态的总体概括可以表述为：

The overall generalization for forming the English past tense can be stated as:

音系理论

Phonological theory

前面几节讨论了人类语言中常见的一些对比和交替。但音系学家想知道的远不止这些。他们想要一个概括，而不仅仅是一个列表。音系学家不想知道“波兰语的声音清单是什么？”而是“任何语言中可能的声音清单是什么？”他们想知道 不仅仅是“不同语言中有哪些常见的变化？”而是“任何语言中都可能有哪些变化？”他们想知道的不仅仅是“俄语和乌克兰语有什么不同？”而是“语言可以有多大的不同？”

The preceding sections have discussed some contrasts and alternations that are common in human languages. But phonologists want to know more than this. They want a generalization, not just a list. Phonologists don’t want to know just “What is the inventory of sounds in Polish?” but “What is a possible inventory in any language?” They want to know not just “What alternations are common across languages?” but “What alternations are possible in any language?” They want to know not just “How are Russian and Ukrainian different?” but “How different can languages be?”

我们已经看到，语言选择不同的语音维度（不同的声道​​手势）来编码它们的对比。发声、送气、方式和发音位置都可以用于对比。同样，我们已经看到这些相同的维度被用来定义与音系交替相关的声音集。 独特特征理论的目的是对语言中所有可用的语音维度进行编码，以编码对比和自然类别。

We have seen that languages choose different phonetic dimensions (different vocal tract gestures) to encode their contrasts. Voicing, aspiration, manner, and place of articulation can all be used contrastively. Similarly, we have seen these same dimensions used to define sets of sounds that are relevant for phonological alternations. Distinctive feature theory aims to encode all the phonetic dimensions that languages have available to encode contrasts and natural classes.

语言学家罗曼·雅各布森 (Roman Jakobson, 1896–1982) 提出，每个相关维度都可以看作是正反对比。说话者选择声音是 [+voice]（有声带振动）还是 [–voice]（无声带振动）；[+nasal]（开放软腭）还是 [–nasal]（闭合软腭）；[+sonorant]（没有气流阻塞的响声）还是 [–sonorant]（有气流阻塞的阻塞声）；[–continuant]（塞音）还是 [+continuant]（不是塞音）等等。雅各布森为他的每个特征提出了声学和发音定义。

The linguist Roman Jakobson (1896–1982) proposed that each relevant dimension could be thought of as a plus or minus contrast. The speaker chooses whether a sound will be [+voice] (with vocal fold vibration) or [–voice] (without vocal fold vibration); [+nasal] (open velum) or [–nasal] (closed velum); [+sonorant] (a sonorant sound, without airstream obstruction) or [–sonorant] (an obstruent sound, with airstream obstruction); [–continuant] (a stop) or [+continuant] (not a stop), etc. Jakobson proposed both an acoustic and an articulatory definition for each of his features.

每个音素都可以根据一组独特的特征来定义：例如，[m] 将是 [+唇音，+鼻音]。特征也可用于定义自然类别以及影响它们的变化。西班牙语中的浊塞音将是 [–sonorant, –continuant, +voice]，而塞音到摩擦音的转变将是 [–continuant] 变成 [+continuant]。

Every phoneme could be defined in terms of a set of distinctive features: [m], for example, would be [+labial, +nasal]. Features could also be used to define natural classes, and the changes that affect them. The class of voiced stops in Spanish would be [–sonorant, –continuant, +voice], and the change from stop to fricative would be [–continuant] becomes [+continuant].

通过提出一组固定的、有限的通用特征，雅各布森试图定义所有可能与音系相关的语音维度——即可用于对比和交替的语音维度。一种可能的人类语言将使用这些特征，而且只使用这些特征。雅各布森提出的许多特征至今仍在使用；其他特征已被替换或改进。

By proposing a fixed, finite set of universal features, Jakobson attempted to define all the phonetic dimensions that could be phonologically relevant – that is, that could be used for contrasts and alternations. A possible human language would use these, and only these, features. Many of the features Jakobson proposed are still in use today; others have been replaced or refined.

音系学家还关注描述音素（底层表征，或 UR——大脑中单词的存储方式）与音位变体（表面表现形式或 SR——单词的实际发音方式）。诺姆·乔姆斯基和莫里斯·哈勒在1968 年出版的颇具影响力的著作《英语的发音模式》中提出，音位变体是通过应用以下方式从音素派生出来的：音系规则。这些规则的形式为

Phonologists are also concerned with describing the relationship between phonemes (the underlying representation, or UR – the way words are stored in the brain) and allophones (the surface representation, or SR – the way words are actually pronounced). Noam Chomsky and Morris Halle in their influential 1968 book, The Sound Pattern of English, proposed that allophones are derived from phonemes by the application of phonological rules. These rules took the form

其中 X 是受影响的声音类别，Y 是变化，A__B 是上下文，全部以区别性特征的形式书写。上述规则可以读作“在 A 和 B 之间的上下文中，X 变为 Y”或“AXB 变为 AYB”。其他特殊字符用于表示单词和词素边界：#___ 表示“词首位置”，___# 表示“词末位置”。您能读懂以下两个描述上述变化的音系规则吗？

where X is the class of sounds affected, Y is the change, and A__B is the context, all written in terms of distinctive features. The rule above can be read “X becomes Y in the context of between A and B” or “AXB becomes AYB.” Other special characters were used for word and morpheme boundaries: #___ means “word-initial position,” and___# means “word-final position.” Can you read the following two phonological rules, which describe alternations discussed above?

第一个规则规定，清塞音出现在词首时会变成送气音。这描述了英语中 [p]、[t] 和 [k] 的词首送气。第二条规则规定，浊塞音跟在另一个不间断音之后时会变成持续音（即擦音）。这描述了西班牙语中浊塞音 [b、d、g]前面有 [ ] 时会变成浊擦音 [ β , ð, ɣ ]。元音。

The first rule states that a voiceless stop becomes aspirated when it is produced at the beginning of a word. This describes word-initial aspiration of [p], [t], and [k] in English. The second rule states that a voiced stop becomes continuant (that is, a fricative) when it follows another nonstop sound. This describes how a voiced stop in Spanish [b, d, g] changes to voiced fricatives [β, ð, ɣ] whenever it is preceded by a vowel.

在一个派生规则将 UR 改为 SR。回想一下上面关于单词grandm发音的讨论。基本、底层的表示形式在中间有三个辅音 [ndm] 序列。首先，通过删除 [d] 简化此序列。然后，[n] 代替发音与接下来的 [m] 同化，现在它与 [m] 相邻。结果是中间辅音实际上发音为 [mm]。派生显示了这一系列变化。因此，[græmmʌðər] 的派生如下所示：

In a derivation, rules change the UR into the SR. Recall the discussion, above, of the pronunciation of the word grandmother. The basic, underlying, representation has a sequence of three consonants [ndm] in the middle. First, this sequence is simplified, by deletion of the [d]. Then, the [n] assimilates in place of articulation to the following [m], to which it is now adjacent. The result is that the medial consonants are in fact pronounced [mm]. A derivation shows this series of changes. Thus, the derivation of [græmmʌðər] looks like this:

音系学家提出了不同的规则编写方法。例如，将同化视为特征共享而不是特征改变可能更有意义。人们可以通过绘制一个辅音与另一个辅音共享的单个特征来编写浊音同化规则：

Phonologists have proposed different ways of writing rules. For example, it might make more sense to think of assimilation as feature sharing rather than feature changing. One might write a rule of voicing assimilation by diagramming a single feature being shared from one consonant to another:

这种编写规则的方式称为自动分段表示，因为这些特征被视为独立于段串的。

This way of writing rules is called autosegmental representation, because the features are treated as autonomous from the string of segments.

最近，音系学家开始质疑音系规则是否是思考底层表征和表层表征之间关系的最佳方式。规则可以描述交替，但它们没有捕捉到交替往往有原因这一事实——使单词符合音位学约束。音系学的最新研究强调了约束在调解底层表征和表层表征方面的重要性。这些研究提出，所有语言都共享同一组约束，但对它们的排序不同。给定形式可能会违反排序较低的约束以满足更重要的约束。例如，将 /grændm ʌ ð ə r/ 发音为 [græmm ʌ ð ə r] 表明，英语更重视“不能有三个连续的辅音”这一约束，而不是“UR 中的每个音段都必须出现在第一个音节中”这一约束。“SR。”

More recently, phonologists have questioned whether phonological rules are the best way to think of the relationship between underlying and surface representations. Rules can describe alternations, but they don’t capture the fact that alternations tend to take place for a reason – to bring words into conformity with phonotactic constraints. Recent work in phonology has emphasized the importance of constraints in mediating between underlying and surface representations. Such work proposes that all languages share the same set of constraints, but rank them differently. A given form may violate a lower-ranked constraint in order to meet a more important one. For instance, the pronunciation of /grændmʌðər/ as [græmmʌðər] indicates that English places a higher priority on the constraint that says “Don’t have three consonants in a row” than on the constraint that says “every segment in the UR must appear in the SR.”

章节摘要

Chapter summary

本章的目的是从语音学家和音系学家的角度描述语音。语音学家研究语言声音的物理方面：声道结构的运动、发音的位置和方式发音、声波在空气中的传播、听觉和言语感知、基频和共振峰结构的计算机测量。音系学家研究声音模式的更抽象的组织：音节结构、音位约束、交替、底层和表面表示之间的关系。

The goal of this chapter has been to describe the sounds of speech, from the point of view of a phonetician and of a phonologist. Phoneticians study the physical aspects of linguistic sounds: Movements of the structures of the vocal tract, place and manner of articulation, the propagation of sound waves through the air, hearing and speech perception, computer measurement of fundamental frequency and formant structure. Phonologists study the more abstract organization of sound patterns: syllable structure, phonotactic constraints, alternations, the relationship between underlying and surface representations.

我们在本章的开头提出了音系学家和语音学家提出的问题，并尝试调查一些已提出的初步答案。音系学家和语音学家已经了解了很多关于语音是如何产生的、如何被感知以及如何组织的。但仍有许多问题尚待解决。

We began this chapter by posing questions that phonologists and phoneticians ask, and have attempted to survey some of the preliminary answers that have been proposed. Phonologists and phoneticians have learned a lot about how speech sounds are made, how they are perceived, and how they are organized. But many questions remain.

• 还有哪些新的声音和声音模式有待发现？
• What new sounds and sound patterns remain to be discovered?
• 我们如何才能最好（并且最快）地描述和保存正在消亡的各种语言的声音和发音模式？
• How can we best (and most quickly) describe and preserve the sounds and sound patterns of the diverse languages that are dying out?
• 随着我们对我们的嘴和耳朵如何工作越来越了解（通过越来越复杂的测量设备），我们如何将这些知识融入我们的声学和​​发音模型中？
• As we gain more and more knowledge of how our mouths and ears work (from more and more sophisticated measuring devices), how can we incorporate this knowledge into our acoustic and articulatory models?
• 随着我们对大脑工作原理的了解越来越多，我们如何将这些知识融入到我们的语音模型中？
• As we gain more and more knowledge of how our brains work, how can we incorporate this knowledge into our phonological models?
• 认知模式和发音事件如何相互影响和制约？
• How do cognitive patterns and articulatory events influence and constrain each other?
• 正确的区别性特征集是什么？语音特征和发音姿势是一回事吗？
• What is the right set of distinctive features? Are phonological features and articulatory gestures one and the same?
• 底层表示和表面表示之间的关系是基于规则的还是基于约束的？
• Is the relationship between underlying and surface representation rule-based or constraint-based?
• 我们如何解释语言变化和语言变异？
• How can we account for language change and language variation?
• 我们如何才能更好地了解儿童如何习得母语的音系和语音？
• How can we better understand how children acquire the phonology and phonetics of their native language?
• 我们如何才能更好地理解学习非母语的过程，并帮助那些正在努力学习非母语的人？
• How can we better understand the process of learning a non-native language, and help those who are struggling to do so?

回答这些问题的研究仍有待完成。

The studies that will answer these questions remain to be written.

练习

Exercises

进一步阅读的建议

Suggestions for further reading

Johnson , Keith 2003，《声学和听觉语音学》，第二版，牛津和马萨诸塞州马尔登：布莱克威尔。这是一本非常易读的关于声音物理学、语音感知和计算机语音处理的入门书。

Kachru、Braj、Kachru、Yamuna 和Nelson、Cecil（编）2009 年，《世界英语手册》，牛津和马萨诸塞州马尔登：布莱克威尔。由专家作者撰写的 42 个章节涵盖了英语国际方言的历史和多样性。

Kenstowicz , Michael 1994,生成音系学, 牛津和马登, 马萨诸塞州: Blackwell。更高级、更全面的音系学介绍，包含大量数据集。

Ladd，DR 2000，《语调音系学》，剑桥大学出版社。从语言学角度介绍语调的研究。

Ladefoged，Peter 和Johnson，Keith 2010，语音学课程，第 6 版，Wadsworth Publishing。语音学实用指南，强调发音和转录练习。

Ladefoged , Peter 和Maddieson , Ian 1996, 《世界语言的声音》 , Oxford and Malden, MA: Blackwell。对 IPA 音标中所有声音的发音和声学进行了调查和详细描述。

Spencer , Andrew 1996，《音系学》，牛津和马登，马萨诸塞州：布莱克威尔。一本非常通俗易懂的入门书。

Wolfram , Walt 和Schilling-Estes , Natalie 2005。《美国英语：方言和变体》，牛津和马萨诸塞州马尔登，布莱克威尔。正如标题所述，这是对美国英语变体的调查。

Yip , Moira 2002, Tone , Cambridge University Press. 全面概述了声调音系学。